goal17 Podcast

You will often hear the idea of “human-in-the-loop” as a core principle of advocacy for AI’s deployment in decision cycles. Those who are worried about AI agents running rampant and making dangerous decisions on our behalf suggest that a key safeguard against bad behaviour - and to maintain accountability - is to ensure that humans have final say in any process.

For those that have been in those loops, however, the distinction starts to seem pretty academic. One of the first areas of knowledge work to experience the application of AI at scale is software development, where, in the last 18 months, coding agents have gone from a highly contested “cheat” to an unavoidable way of working. For those that haven’t used a coding agent, the transaction goes something like this: you describe the change you’d like to make, either immediately or in “planning mode” (which is meant to trigger a back-and-forth to plan the details of the approach) and then the agent gets to work making the changes. Early on, the planning mode was often a very extensive exchange on the details. What would ensue would be a rapid burst of activity, then a pause for approval showing the command that was about to be executed, and a Yes/No/Say something else menu. There might be seconds between approval steps, or minutes. If you had multiple sessions going, you would be quickly toggling between windows, hitting Enter before toggling to the next window needing your attention.

Goal17 is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

Strictly speaking, this kept you in the loop. With a number of parallel sessions going, however, it became an exhausting task of juggling multiple contexts and threads while trying to maintain enough focus to catch when things were going wrong.

Fast forward to now, and “Planning Mode” now generally consists of two or three questions back from the model, then approval to go into “Auto Mode”, where the agent will go off and execute large chunks of work before looping back for input. While there are restraints to keep it from executing the most destructive commands on its own, the agents now have considerable latitude to act without input.

What many have experienced when being the human in the loop in this type of work is a kind of numbing, a cognitive fatigue that sets in where the constant bombardment of micro-decisions wears down your ability to apply judgement, making the user more and more compliant.

This will, I think, become a very important design consideration in the next wave of technology systems that employ artificial intelligence. We’ve seen already with self-driving cars that partial accountability for humans just does not seem to be a workable model. In a B.F. Skinner-eque case of conditioning, each time the machine demonstrates sound ability, the operator is conditioned to trust it more and more, until their oversight becomes ceremonial. Asking a driver to sit at attention at the wheel for extended periods, waiting for the moment the system - perhaps after hours of driving - makes an error in judgement sets a bizarre expectation of almost monastic focus. We already design curves in long roads to force driver attention - why we think human attention has changed since then is a mystery to me.

For any systems that are meant to interface with human decisions, this should become a central design consideration: where are human decisions most needed, and what pattern of cognitive engagement is required to prepare them to be able to fully engage in those moments?

Put differently, how do we design AI systems not as a replacement of critical cognitive tasks, but as an augmentation of them?

For decision systems, that means understanding how people engage with each part of a work process, and what engagement with each part of that process means for their own relationship with the decision to be made.

Systems that take this into consideration can become a true prosthetic for human cognition - systems that allow us to extend our own capabilities by augmenting parts of a process where our faculties might fail us, while facilitating us in areas where our experience, intuition and responsibility should be centred.

The fact that the dominant interface for interacting with many AI systems should be a signal of what most AI companies believe our relationship with the technology should be. While there are recent cases of major decisions being made through a chat, they are generally not held up as examples of great decision-making. If I wanted to have a substantive conversation with someone, texting is not how I would do it. In the case of interacting with a near-omniscient machine, the chat interface is inherently disempowering: ask me a question, I’ll give you the answer. While it has its advantages, long exchanges tend to funnel the user towards a result, with the thought process lost in a long string of text responses among a long list of chats. Trying to shape anything more comprehensive is not a part of the interface - you are left asking for new versions of whatever the chatbot produces.

In designing KnowledgeCommons.ai, which is designed to assist in research and decision making for individuals and groups, that has resulted in three broad areas of capability:

* Capture: how can all the relevant information be brought into a single, flexible space?

* Analysis: how can that information be “played with” to allow experimentation, iteration and thought?

* Feedback: how can emerging patterns be fed back to an individual or group to increase understanding and shape the direction of their inquiry?

For decision-making, I think the analysis and feedback components become the most interesting parts of the system. I know from using various note taking apps over the years, from Notion to Zoom, to any others, the problem quickly became that I was inundated with meeting summaries, which I didn’t know what to do with. Sorting them was a pain, then referring back to them was worse. For someone with a packed schedule of meetings, the backlog quickly became worse than having no notes at all. Added to that, taking notes in meetings by hand was a way of training my attention and capturing thoughts as I went. Replacing that light cognitive filtering not only lost the original benefits of running meetings through my own thought process, but it replaced it with a product that was overwhelming and confusing.

Further, every meeting would likely have a bunch of other inputs and outputs - documents people were referencing, links they were sharing - that all should be a part of the context, but when trying to come back to the topic later were very difficult to reassemble.

So “analysis”, for me, in the course of work, meant a constant struggle to reassemble all the pieces that were meant to be a part of a certain discussion or decision. Documents scattered between systems, links, conversations and notes, none of which connect to each other, with only ad hoc methods of bringing them back into something I could work with.

In practical terms, that meant, for me, an interface where I could play with information at a granular level, from wherever it came from, in order to remix it into something that made sense to me. I’ve always thought well on whiteboards and with sticky notes, for example, but always found it cumbersome to transfer information into and out of those formats. So one of the interfaces I built was a canvas where I could “explode” long documents into sticky notes, so I could gather and sort some of the concepts into a broader context. I also built a “storyboard”, that allowed me to gather ideas that popped up in documents, meetings or websites, and drag them into pages that would then become documents - all keeping reference to the original provenance of each idea.

In both cases, the idea is to make all of the overwhelming input of a system that can “capture everything” into a surface that allows for consideration and deliberation, while not trapping that information in a format that prevents you from moving along in the collaborative process. A simple toggle turns the storyboard into a document that you can continue to work on, now that the larger ideas are laid out. Surveys can flow their results into the same interface, or, conversely, everything on the canvas can flow out again as a spreadsheet.

AI, in this case, is being relegated to tasks of gathering, collating and preparing elements for human consideration beyond a simple Yes/No prompt.

My son, who is in his first year in university and is using the system to keep track of notes and research for all his classes, suggested a feature that has become a core part of the model. “I wish I could highlight things,” he told me after his first week of using it. Highlighting, now, can be done across the whole platform - reading a summary or an input document, you can highlight any passage, then gather all your highlights into a spreadsheet or canvas, or, if you’re trying to create a report, into a storyboard.

This, to me, is about trying to find the right relationship between humans and AI so that we are supporting thought, not replacing it. Thinking requires the right kind of friction - enough work of exploring and combining to create something new - while eliminating unproductive friction; file formats, scattered folders and an inability to find that perfect quote that “I saw somewhere.”

For anyone that’s been watching this space, you might have noticed that I’ve been rather quiet for a while. Some of my last posts were exploring the impacts and uses of AI in consulting, collaboration and public engagement.

My early experiments quickly convinced me that AI tools offered significant value in collaborative work but I didn’t see a way to keep it as a core tool in my practice with the options that were available. What I did see, though, was that with the right tools, we could finally overcome some long-standing challenges in collaborative work to enable new forms of collective action and complex decision-making.

Goal17 is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

Through several decades of work in designing large-scale, complex, multi-stakeholder decision-making processes, there were several persistent problems we’ve been working to correct for that finally seemed within reach, and, to me, offered the possibility for a real leap forward in how we organize ourselves around complex decisions.

Whether in large organizations, in projects or across coalitions, the quality of coordination and decision-making has always been a function of how quickly information gets where it needs to be, how widely distributed it is, and whether those in authority are aligned in their understanding.

What became clear very quickly was that we could improve the speed of information passing through teams and networks, ensure that “edge insights” become more distributed and bring groups to shared understanding more efficiently. As I tested recording in breakouts and across sessions, I could see that assumptions we had held about the quality of what groups would report about their work paled in comparison to the richness of complete capture.

I felt that if you could synthesize disjointed, distributed or parallel work quickly enough, at scale, that this would no longer be a question of documentation after the fact, it would completely change the speed and comprehensiveness of collaborative work.

Consultations would be able to not only capture in minute detail the perspectives of constituents, but could adapt their focus - at scale - as unexpected themes emerged.

Just off the top of my head:

* Strategic conversations could merge the views across all stakeholders and document inputs.

* Workshops could iterate quickly on emerging ideas and save time on lengthy reports between teams.

* Conferences could become knowledge-producing hubs.

* “Systemic” conversations and policy processes could assemble first-hand knowledge from more people than was ever practical before.

I was convinced of the potential, and I remain convinced. Anyone in professional services that is involved with engaging and assembling knowledge, perspectives and intent across large groups will, I believe, fall behind if they don’t account for this step-change in capability. However, I am not a blind AI fanboy. Ethics and modelling outcomes have been central to my practice for years, and I have been a long-time critic of how the unprincipled development and application of technology has had predictably dire outcomes for society. Having previously delivered AI ethics and governance projects, I also knew that rolling out something like this would require making decisions with very real consequences.

I am not going to go into detail here on how, exactly, I addressed each of these issues. Many are worthy of future posts. This is not even an exhaustive list. Suffice it to say, however, that in my view, AI platforms and any prospective users of those platforms, need answers to these questions and many more. During development, I’ve noticed a deep chasm between those with technical expertise and those who are shaping policy and engaging in discussions on ethics. Part of my motivation in developing this platform has been to create a kind of “thesis” project that is fully production-ready, that is a technical reflection of many of the more philosophical, ethical questions relating to this technology. It’s why my writing has lapsed in this space for a while: I was writing something else.

Now on to some of the issues and questions I addressed in my design.

Privacy

Listening to conversations at scale creates a host of privacy concerns. If “collective intelligence” is the goal, then trust and the confidence of individuals to speak freely is a precondition. If they are suspicious of the recorder, they won’t share.

* What happens with the recordings, and who can listen to them?

* How will personal side conversations be managed?

* Where are the recordings being processed, and what are the policies of the processor?

* Are statements being personally attributed, and if so, who has access to those attributed statements?

* Are my conversations being used to train models, build someone else’s product, or serve someone else’s secondary business model?

Security

If these conversations are confidential, proprietary, or otherwise sensitive, there are a host of questions that emerge end-to-end in terms of how data is gathered, transmitted, processed and stored.

* Can the radio frequencies from any recording devices be intercepted in order to eavesdrop on in-person conversations?

* Do recordings stay on any recording media between or after meetings?

* If recordings are being transmitted, can the data be intercepted in transit?

* Can AI models learn, retain or leak any of the materials being processed beyond the scope of a given group?

* Is all original, processed or derived data stored securely?

Cognition

Artificial Intelligence, when applied to cognitive work, can either augment human performance or systematically degrade it. Especially when applied to difficult, strategic conversations about the unknown, AI can adversely impact the ability of humans to tackle thought-work.

* Is AI being used to replace creativity, decisions or tasks requiring judgement?

* Who is generating recommendations or options - humans or AI?

* Is the “work” being done by the AI assistive, or authoritative?

* Does the application of AI generate additional cognitive load through over-generation?

Human Agency

AI systems can be extractive and erode the importance of human roles, or can be assistive and extend human capability.

* Is the role of AI generating new value in the work being done by humans, or replacing existing roles previously done by humans?

* Does AI draw from outside or unverifiable sources that compete with qualifying knowledge and expertise of human operators?

* Is AI being used to make judgements that carry liability or normally demand formal accountability?

* Is the role of AI creating dependence that might diminish the critical faculties of human users of the system over time?

* Who decides what is important, and are those decisions clearly and transparently recorded?

Sovereignty

The development of AI is taking place on the backdrop of renewed geopolitical conflict, and control of the technology, infrastructure and data involved is now subject to state surveillance, restriction and control.

* Is any part of the AI system - from collection, to transmission, processing and storage - transiting a potentially adversarial jurisdiction?

* Are any critical components of the system vulnerable to import or access controls that might render the system inoperable?

* Does the system rely on the goodwill of foreign or adversarial governments?

* Is any part of the system in a jurisdiction that can compel involuntary access or disclosure?

* Is the system in a jurisdiction that will credibly enforce stated privacy and content protections?

Trust

Many of the leading AI labs have systematically violated existing copyright laws and are backed by governments that prioritize rapid growth and technological supremacy over human, labour and privacy protections. Further, frontier labs have opaque cost and pricing structures that do not appear to align with publicly available estimations on delivery cost.

* Does the AI provider have a verifiable set of protections in place for secondary uses of customer data?

* Has the AI provider ever been found to infringe on the use of data they do not have rights to?

* Do the AI provider’s prices accurately reflect the cost of delivering services, such that current prices are a reliable predictor of future prices?

* Has the provider ever been seen to misuse, mishandle or misappropriate private information?

* What assurances do users have on the accuracy or provenance of information provided by the service?

* Are the stated aims and objectives of the provider or its leadership aligned with the interests of its customers?

And So…

Where did this leave me? There was, somehow, nothing I could find on the market that could record 20+ conversations in a room simultaneously, and no provider for processing all of that information in a way that satisfied all of the conditions above (and many more).

So, that is why I built Knowledge Commons and the Raven Recorders. I could sit on the sidelines and complain about how the AI industry was taking the wrong path, or, I could build something that demonstrated what I thought the right path should look like.

Goal17 is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

A caveat up front: The Value Web has always been a community effort, and these are just my personal reflections as we reach a new part of the journey.

I was proud to be a part of the Value Web when it was a scrappy, international collective of practitioners that had come together to apply their talents to tackling some of the world's most intractable problems. I was a part of the board when we changed the mission statement to "transforming decision-making for the common good". It felt right. It felt big. It felt like it mattered.

But things have changed. And if you haven't noticed, decision-making doesn't seem to be doing too well right now.

What I have come to believe, however, is that we have, among us, the tools that we need to make a fundamental shift in how we approach the problems of our times.

Goal17 is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

Where We Came From

This all started out - at least for me - as a corporate thing. Capgemini Consulting, by way of Ernst and Young, had acquired a methodology from a small, obscure and boutique group called MGTaylor for a facilitated process of group decision-making.

I was, perhaps, too inexperienced to fully understand why we were doing what we were doing when I started...but I knew something felt right about it. In the crudest sense, as I understood it then, what we were doing was creating and facilitating a process by which a group could openly and collectively evaluate the possibilities of potential future strategies and collaboratively find a path forward together.

Using a clear method for process design along with a set of design principles and concepts for the creation of physical environments to foster collaboration, the method achieved significant scale across the world and spawned a global community of practitioners dedicated to supporting collective decision-making. I'm abbreviating rather substantially in this history, of course.

When I first got involved, what stood out for me was the openness and democratic nature of the process. I thrived on that. From the moment I entered the workforce, I wondered why good ideas mattered less than hierarchy; encountering a method for allowing the best ideas from a group to emerge - no matter who they were from - was a breath of fresh air to me.

I didn't really understand the real meaning of what we were doing until years later. It was enough for me that the work we were doing seemed to give meaning to groups within various companies that were tackling their own challenges. They seemed to gain inspiration and energy simply by being engaged in conversation about how to approach the challenges facing their companies.

And Then, The Value Web

Over time, I came to genuinely appreciate the work we were doing. Participants in our processes became deeply engaged in the difficult work of navigating what were - oftentimes - existential threats to the organizations they were working for. I learned that by designing a collective process around how people think, which is often messy and non-linear, groups would build ownership and intent around their work that I hadn't seen otherwise in the working world. When they were meaningfully engaged, they leaned in.

And there was method around all this. We would often say "trust the process"...and we would mean it, because the "process" would reliably produce results.

When I came across The Value Web, they were doing something that I found very interesting: they were applying the methods we had used in our consulting context for collective decision-making in non-corporate settings, most interestingly in settings where nobody in the group was from the same organization.

I can't stress how important this factor was in the evolution of our thinking and our work.

Collective decision-making and co-creation is comparatively easy when everyone in the process is obligated to be there, has the same interests in the outcomes, and might be fired if they don't meaningfully contribute. It's participatory, but with consequences. And the smart participants know what game is being played.

Collective decision-making when everyone is from a different organization, when they do not share accountabilities and when ownership of the outcomes is unclear, is a different animal entirely. This was the arena I found the Value Web playing in. Diverse groups from multiple sectors and segments of society trying to figure out solutions to intractable problems. Together.

What became abundantly clear was that there was a real gap in how to balance the inclusiveness required to involve all the necessary stakeholders with the decisiveness required to move things forward. And what we were doing appeared to be working.

What We Did and What We Learned

For years, we operated as a collective. Awkwardly. Somehow, we, as a group, found organizations to work with that needed support, and we created beautiful, immersive and transformative experiences for leaders facing critical challenges.

We worked with the World Economic Forum to reimagine its gatherings. We worked with UN agencies to find points of collaboration between agencies to tackle complex challenges. We tackled projects on climate, nature, public health, resource scarcity while also learning the fundamental principles of community design and coalition-building.

More than anything, I think that what we learned was that while having methods to productively and decisively engage individuals in big decisions was useful in large organizations, it was fundamental to making progress in settings where the stakeholders weren't beholden to the same "boss" but, nonetheless, had common stakes in a problem that none of them owned individually, but all of them were responsible for collectively.

Over time, we reflected on what we were doing and realized that in our efforts, there was something bigger at play.

Shared Intent and Collective Intelligence

It turns out that having a global community of people obsessing over how decisions get made results in some fairly significant insights. Over hundreds of projects, there was a very real validation that all of the factors surrounding HOW decisions get made are as significant as the decisions themselves.

By focusing on the human experience of collaboration, the emotional journey, the heuristics and shortcuts of human cognition and the labor of human trust and connection - all of which were considered unprofessional, irrelevant externalities in traditional decision-making methods - we were able to create deep and stable transformations of the groups we worked with.

We came to see that work in these systemic contexts focused around three design challenges - distributed intelligence, individual action and personal intent.

It was similar to the work we did in the corporate context, but exploded to a scale that required us to extend the tools and models we used.

At its heart, though, the problems were deeply human. Intent was everything. With loose ties, individual intent that became shared intent was the most potent element in driving change. And collective intelligence simply meant that with the size of the challenges, no individual fully understood every part of the problem, so a meaningful process to allow members of a group to come to shared knowledge based on collective input was the most reliable way to ensure that decisions were based on the best information and could account for the many potential consequences.

Through it all, we validated that a model-driven design process helped create structure in otherwise confusing and unruly circumstances, because the common denominator, regardless of industry or domain, was the human condition.

Reaching the Limits of a Model

To achieve all this, The Value Web walked an organizational tightrope for many years. As a "collective", it was made up of a group of more than 30 practitioners delivering work together under a common name. Most of those members either operated another company or worked somewhere else, and the common brand was used as a neutral space to collaborate on projects for the common good.

It always felt somewhat temporary and incomplete. It had enough structure that we could work together, but each time we attempted to formalize it, the changes risked upsetting the balance that allowed a group of people who might otherwise be competitors to work together. The energy - our shared intent - was always in delivering meaningful work together to try and make a difference, and that energy was always tested when we tried to evolve the structure.

We had validated that well designed, effectively supported decision-making processes could make a difference at the very highest levels, and with the most difficult and complex problems, but we had done so using a structure held together with chewing gum and duct tape.

And then we noticed a set of challenges emerging that caused us to re-evaluate the path forward.

First, the rise of Design Thinking muddied any kind of comprehensive understanding of deeper methods. The runaway popularity of the set of techniques around Design Thinking made it more difficult to articulate the important nuance of designing thinking. And our community was small, obscure, and not widely known.

Second, our extended community of practice was getting older. Although practitioners of our craft had, collectively and individually, achieved a remarkable degree of success, the obscurity that had always given it an edge now acted as an impediment to a generation of people who didn't even know these practices existed. Extinction didn't seem out of the realm of possibility.

Thirdly, our own practice could not evolve if it was not clearly defined enough to enter into conversation with other practices. There was no clear frame of reference outside of our community for what on earth we were doing.

Finally, and most importantly, there was no conceivable way that we could achieve our mission of transforming decision-making for the common good simply through scaling our service delivery. We could never grow enough or deliver enough projects to meaningfully change the broken processes that were steering our species off a cliff.

So the question became, given all that we've learned, how could we, as a community, change societal expectations of what constitutes "due process" in how critical decisions are made?

Ending the Old to Create the New

Svenja Ruger, Tanja Kerlo and myself took on the project of imagining what that might look like. First, the organization could no longer deliver projects itself, so that it would not be in conflict or competing with its own members. This was the most emotionally difficult part of the process, as we all had a connection to this dysfunctional but beautiful way of working together.

We came up with a minimal structure that would allow us to conduct research, establish the boundaries of a profession, and provide a learning path for new practitioners and to advocate for leading methods in complex decision-making. We call it Process Activism.

The most important element, however, was that it would take the time, intelligence and talents of our entire existing community of practice to come together and bring this into the world. With thoughtful practitioners spread all across the world choosing to move this forward together, it might just stand a chance of really changing things for the better. With this community as a start, and with an intent to learn with other communities of practice, it would be conceivable that collectively we could shift towards more inclusive, informed and robust collaborative decision-making.

We believe that we need to build community around key challenges, so there is a shared understanding among stakeholders not around what a given decision should be, but how we should be approaching those decisions.

One week ago, we made our case to 100 members of our community that had gathered for The Happening. We relinquished control of the organization to allow its rebirth as a community project.

What I witnessed was a group of people stand up and answer the challenge. What I felt in that room was the emergence of shared intent, and a genuine desire to build something together to ensure that we, as a species, can work towards a time when we can say that all decisions that affect our future were made in the best interests of all, and represented the very best of our knowledge and abilities.

Matt and Gail Taylor gave an incredible gift to this community when they assembled and nurtured this method. This community has now expressed the desire to honour what we’ve learned by extending it, sharing it and allowing something new to grow.

I am very excited to work together with this amazing community to create a future where we can not only say our decisions reflect our collective intelligence, but that, together, we have nurtured the ability to act with collective wisdom.

You should join us.

Context

In the midst of all that is going on in the world, Canada just had a federal election. Normally, Canadian politics isn’t something that gets the blood racing, but these are not normal times, and this was not a normal election.

In the course of my work, I find I am more routinely keeping an eye on the flow of global politics, but over the last couple of years I have started to focus a lot more on the threats facing Western democracies, both from the corrosive effects of digital platforms but also with the increasing intensity and impact of influence campaigns waged by autocratic states. I was getting increasingly worried that the underpinnings of our democracies were crumbling, and we seemed ill equipped to counter the challenges we were facing, and as a result, I tried to focus my work, where I could, on some areas that could contribute to our collective defence.

All the while, the Canadian government was staggering along, with a beleaguered administration that just never quite got its stride again after the pandemic. We have a politician here leading the opposition who had made denigrating Canada into a full time, 2 year project, baiting the increasingly unpopular Prime Minister and drilling into Canadians that our country was broken, despite our better-than-the-global-average recovery from the pandemic, global supply shocks and inflation. A lot of Canadians came to believe him, and even among those of us who didn’t, there was little enthusiasm to support a Prime Minister that seemed to be holding on long past his due date.

And the polls were grim. Support for the Liberal government was at historic lows, and while they held on in Parliament with the support of a coalition party, survey after survey showed that the next election would be an extinction-level event for the ruling party, with a crushing majority for the Conservatives forecast whenever the writ might drop.

Goal17 is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

Honestly, for my part, while I found the Conservative remedies completely unconvincing, I was becoming somewhat resigned to the fact that they would take power, and I felt that, perhaps, it would be best for them to have a term in power if only to prove that they couldn’t just wave a wand to make the world’s problems disappear. I started to think it might be worth them winning just to prove to Canadians that the simple platitudes they were offering to fix our “broken” country had no substance to them, so we could move on.

But then, a few things happened. Trump got elected. The joke about making Canada the 51st state stopped being a joke. Vance travelled to Europe to tell the German military it should be okay with fascism. The US started slapping tariffs on Canada. And then the Liberal government experienced a rapid, unscheduled disassembly.

Suddenly the copy/paste of Trump’s talking points into the Canadian Conservative leader’s speeches felt like less like a sign of admiration, and more like a Manchurian candidate.

I decided on the night that J.D. Vance spoke at the Munich Security Conference that I wanted to be involved in some small way to contribute. The challenge was that, as a resident of downtown Toronto, any campaign door-knocking I could do would be preaching to the converted. I was aware, however, that the misinformation war was unfolding online.

The Plan

My partner, Beth - who has worked extensively in social media strategy - convinced me that if I wanted to reach a lot of people in a short amount of time, the only choice was to try TikTok.

Beth pointed me towards several accounts that had begun focusing on political content as the election drew near and had grown rapidly and achieved significant reach in very little time, which suggested it would provide considerably more reach than knocking on the doors of my liberal neighbours.

To say I was skeptical would be an understatement. In my view, TikTok was a democratic destabilization machine controlled by the Chinese state, which didn’t really make it a great candidate.

The alternatives, however, were not great.

The Platforms

The social media landscape in 2025 is a dumpster fire. No matter what reason you have for using social media these days, you are probably unsatisfied with the experience, and are most likely generally worse off for using it, whether that be for personal reasons or professional.

I had been experimenting with Substack for some time, and putting an ungodly amount of effort into researching, writing and recording posts. While the platform benefits from having great tools for writers and isn’t burdened (yet) with advertisements, I found myself topping out at around 130 subscribers, with every new subscriber a hard fought battle. It seemed to me that it was a great place to bring an audience, but not a platform where you could easily build an audience.

I decided early on that I would try to use my professional network on LinkedIn to try and direct people over to Substack, by posting about my new articles there. It was only then that I realized how hollowed out LinkedIn had become. Firstly, because LinkedIn tries to encourage a posting frequency in their algorithm that is unsustainable for thoughtful production, it has become a hellscape of self-serving humble bragging, with only the rare post rewarding the reader with any actual insight or value. The worst part is that we all seem to know it. Professionally, we know we should at least appear to engage, so there are a smattering of likes and performative comments, and nothing more. The engagement on my posts linking to my articles was shockingly low, with numbers that were only a fraction of the number of connections I had. But worse than that, the “click-through” rate was so low that at first I thought Substack’s analytics were lying to me. On a LinkedIn post with a decent number of likes and even a few comments on the topic, Substack’s analytics would show that almost no one had actually clicked the link.

Instagram doesn’t even seem to know what it is as a platform any more. In response to TikTok’s onslaught of content from people you don’t know, Instagram threw away what was previously its insurmountable competitive advantage: its social graph. Combined with aggressive efforts at monetization by Meta, it is simply a platform for scrolling through advertisements, interspersed with cross-posted TikTok videos from people you don’t know, with bubbles at the top where the few friends you have active on the platform post coffee pictures and conspiracy theories.

Twitter, which now has a name you can’t start a sentence with, is only useful for finding out what its owner is doing, and what other white supremacists think about what he’s doing, is totally unfit for any sustained effort, besides being harmful for your mental wellbeing.

I stopped using Facebook during the pandemic, when it proved to be ground zero for radicalizing its users and turning them against vaccines and democracy. After being attacked mercilessly by a mob after I suggested to an acquaintance that vaccines didn’t cause autism or allow Bill Gates to track us, I decided I was done on that platform. Zuckerberg’s decision to end fact-checking and ban actual news sources from the platform sealed the deal for me.

I had never really used TikTok, mostly because the combination of an addictive algorithm, its ability to “understand” you at a deep level and its connections with the Chinese state had always been incredibly problematic for me, but also, simply, that I didn’t believe it to be a platform where any serious content could exist.

Also, in the context of Trump’s announcement that annexation of my country was on the table, it was worth considering who owned each of the platforms and what their agendas were. All of the platforms were owned by adversarial governments. The CEOs of Meta, X and TikTok all attended Trump’s inauguration, and given Trump’s fixation with Canada and China’s ongoing feud with the Liberal government and documented attempts to interfere in Canadian democracy, one had to assume that there could be interference in political discourse on all of these platforms.

Finally, while I had friends and family on Instagram and Facebook (100-200 connections), an old Twitter account with about 150 followers and a LinkedIn network of around 1500 connections, I had exactly zero followers on TikTok, as I would be setting up an account for the first time.

But, I was determined. No matter how small the contribution to the discourse, I wanted to do something, even if it was only correcting some misinformation, to help in the election.

Oh, and one other note; I was, and am, fully aware that maybe the problem with engagement on the things I’ve written wasn’t an algorithmic problem or a platform problem. It was also possible that I’m just boring.

Playing to the Algorithm

Beth laid out a simple formula for me. She was adamant that if I followed the formula, I would see results and the algorithm would respond, but if I didn’t, and I deviated, or slacked off, the algorithm would be merciless.

She also suggested that because there was so much attention on the election, that the time was now: if I harnessed a national conversation in the moment, the impact would be multiplied.

The formula was simple: you need to post three videos per day, every day, connecting with issues and topics as they arise. You need to find the hashtags for your topics, and respond to every early comment on your posts as they come in, while posting comments on the posts of a few, related, creators around the same time that you post your own videos.

While this was obviously a difficult pace to maintain, I was determined to give it a shot.

Given my experiences on other platforms, my expectations of a new platform with zero followers were pretty low. My very first post, however, got over 300 views almost immediately. Coming from LinkedIn, I thought that was pretty good. My second post, later that day, got 2,500. I was flabbergasted. I had only gained 1 follower from the first post, so that caught me off guard. The numbers seemed to fluctuate up and down, but a few more posts in and I had another hit 3,300. Even getting view counts in the hundreds seemed pretty good to me. Here on Substack, where a single post takes me hours of effort, getting above 200 views is a good day for me. Things seemed to be going pretty well, and with every new post, I would pick up a few new followers, something that on this platform I was finding pretty difficult to do.

What shocked me from my early interactions on the platform was the level of engagement from the people that use it. The dynamic was fundamentally different from anything I’d seen on other platforms. Beth would roll her eyes, because, of course, she’d been telling me this for a very long time. But not only were people viewing political content, they were debating it, sharing, and shaping it. The comments would fill up and users would go back and forth on the issues. And these weren’t the usual LinkedIn comments of “Totally agree: ask me how I can help with this using my special formula!” but actual, genuine comments and engagement. I hated that this platform was where the “town square” had moved, but I couldn’t deny that this seemed to be where Canadians were coming to find out what was going on and to discuss it. Love it or hate it, this was where it seemed to be happening.

That’s when I came across Common Sense Carl and some of the conspiracy content swirling around the platform.

The Misinformation problem

What became apparent from the comments was that while this election, for some, was about trying to find a candidate that was best suited to guide Canada in a time of geopolitical and economic turmoil, for many others, it was about something completely different.

To summarize, there were many who believed that Justin Trudeau, and now his carbon-copy replacement, Mark Carney, were puppets of a globalist cabal led by Klaus Schwab of the World Economic Forum to confine us in our Hunger Games inspired 15-minute City districts while being tracked by microchips injected by Bill Gates and the WHO during an imaginary Plandemic that was concocted to break our spirits into submitting to the communist, totalitarian Liberal government that was using carbon taxes to crush Alberta by hobbling its oil industry, all while extracting their hard-earned oil money to fund the profligacy of a corrupt capital in Ottawa and a lazy, welfare province of Quebec. Or something like that. But it definitely, definitely, could all be traced back to the WEF and their master plan.

Having worked as a consultant and collaborator with the WEF for close to 15 years, I felt the need to set the record straight. I realized that, perhaps, the value I could add in this debate was to offer the perspective of someone who had actually worked with the WEF and the WHO, and hadn’t just heard about them on Facebook. So I made a short video explaining what they actually did, and why the rumours had no basis in reality, gave it the title “The WEF Conspiracy” and posted it.

The video got about 40 views, and stopped dead. The normal curve of engagement for even my least engaging videos didn’t emerge. It just stopped.

After some investigation, and using a tool they have for scanning your content, I realized that I had been flagged for misinformation. Being flagged means that your post is no longer “eligible for promotion on the for you page”, which means that the magic algorithm that makes TikTok run will ignore your content. It won’t delete it, it will just ignore it. I tried changing the text description, the hashtags, and re-uploading it, with no luck. Any mention of globalism, globalist, WEF, World Economic Forum or a host of other words would get you suppressed if you tried to debunk anything.

To add insult to injury, when I looked at TikTok’s “creator insights” tool, which suggests trending topics to act as inspiration for creators, along with scripts to follow and related videos. The script was full of false information, and the suggested videos spoke freely about the WEF and all of the conspiracy theories, with view counts in the hundreds of thousands, meaning that they were using the same keywords while still being promoted widely by the algorithm.

This was one of many moments where you could feel the hand of foreign interference tipping the scales. Want to spread division? No problem. Want to stop it? That’s misinformation.

Where the misinformation really thrived was in the comment sections, and pushing back on the increasing waves of crazy comments became a full-time obsession. Some of it was just crazy, but a lot of it appeared to be bad faith parroting of misleading talking points.

There were definitely bots at play, but I think what’s more concerning than that is that this method of consuming information seems to turn people into bots. People with only cursory knowledge of topics post short, punchy videos putting a spin on headlines that are taken out of context, which are then repeated robotically by viewers who take them as truth. You could spot them immediately, because the talking points would be recycled so consistently: any mention of WEF, WHO, no-new-pipelines, equalization payments, 15 minute cities, Jeffrey Epstein, globalists, Liberal corruption, Brookfield, Klaus Schwab, scamdemic, Jared Kushner a lost decade or a host of other scripted points let you know where a commenter was getting their information. Most of these things took no more than a 2 minute Google search to see they were false, but this was evidently a Google search too far for most.

What was most disturbing for me was the interplay between the Conservative candidate, Pierre Poilievre, and the online misinformation. Poilievre would consistently make references to the online conspiracies in ways that he knew would engage those who knew, while often flying under the radar of those who didn’t. This reinforcement by a national party leader only served to reinforce their credibility, and cement the loyalty of a political base that saw him as the only candidate speaking the truth.

The Polarization Problem

I quickly amassed a small but loyal following of a few hundred - which, again, was astonishing given my experience on other platforms - but the reach of the posts I made, thanks to the algorithm, would extend far beyond that. My posts would consistently get hundreds of views, with some intermittently spiking into the thousands.

At a certain point, I noticed that I was clearly being served up to those that the algorithm knew would disagree with my content, while at the same time being served content myself that I would consider inflammatory.

One day when I expressed my frustration to Beth that I was being served “crazy stuff” that the algorithm would have to know I wouldn’t like, she reminded me that that was precisely the point; disagreement is one of the strongest forms of engagement. This clearly was having the effect of increasing the levels of polarization among users, as you would be comfortably surrounded by opinions you agree with, then suddenly served only the most extreme examples from the political opposite.

This polarization could be seen in the aftermath of the election results; following the Liberal victory, many Conservatives had difficulty accepting that the election results were valid, because they hadn’t come across anyone online who agreed with the Liberals, except for the occasional case of the most unhinged leftist. How could the Liberals have won if EVERYONE said they were going to vote Conservative?

The Fact Checkers and the Trolls

There were some creators I came across that put in an incredible amount of work to counteract the waves of disinformation that poured out on a daily basis. One in particular, Rachel Gilmore, was doing near forensic-level investigation of some of the trickier pieces of disinformation throughout the election. One claim that stood out for me was a picture of Prime Minister Mark Carney with members of a pro-Beijing business lobby, with the assertion that he had an in-depth meeting with them, despite his protestation that no such meeting had happened. By piecing together a timeline, and photos from the same event (based on backgrounds, clothing and flooring, no less), Gilmour showed that Conservative accusations were baseless, and the photo had simply been taken at a large political event, sandwiched between photos with other attendees.

The cost for fact-checkers like Gilmour, however, were substantial. Female voices online are particularly targeted, and Gilmour amassed an army of trolls that would follow her to any online space she would go to, culminating in the withdrawal of an offer to do an on-air segment with Canadian broadcaster CTV, which didn’t want to deal with the hassle of right-wing trolls and extremists.

This level of organized harassment should be deeply concerning for anyone interested in having effective political coverage throughout our elections.

A Side Note on China

Remember the part about TikTok being controlled by China? While I am going to go through this in detail in a separate post, it is worth mentioning that my first “viral” posts that got above 10k were about…China. Their performance was suspiciously out of step with any of my other content up to that point. Nestled between two other posts that garnered my then-typical view-counts of 450-650 views, my two posts about China attracted just below 13k and 10k…with the second post being an incredibly lengthy and dry rehash of China’s brushes with colonial powers through history. Coincidence? When I saw the other pro-China content being boosted in my feed at the time, when Trump was rattling his sabre, it was different enough to catch my attention.

Final Results

My original intent was to test whether using an online platform like TikTok would be more effective that volunteering in the campaign to knock on doors in Toronto. Just based on the limits of physical movement, there is likely an upper limit on the number of doors I could have knocked on, and again, it would have been in an area that was already heavily leaning Liberal.

By contrast, in the four weeks leading up to the election and the two weeks following, when I have been countering misinformation on the validity of the election results, here is a snapshot of my reach, having started an account from zero.

What strikes me is not only has the content gathered nearly a million views, it also attracted over 24,000 comments. Twenty. Four. Thousand. On a six-week-old account.

Further, as a test, at a few points during the campaign, I would post the same video both to Substack and to TikTok at exactly the same time in order to see if it was the content itself that was different, of if it was the platforms that made a difference. Here is a side-by-side comparison of the performance between the two after they’d been posted for the same amount of time:

The lesson that I take from this, and for the very few people who read this Substack, that I hope you take as well, is that this, very much, seems to be where the conversation is happening. My account was not even close to being one of the more popular political accounts during the election - those would have all have significant multiples of engagement beyond this. As of today, this account has around 3,000 followers, whereas the other creators leading the discourse all had tens or hundreds of thousands of followers.

And this really matters. I was very uncomfortable to hear, partway through the election, in comments from a large number of people, that my content had become a go-to for election information and a reality-check on what was real and what wasn’t. While this was, of course, my intent; to offer a balanced voice and some sound analysis into the public discussion, it made me realize that for every one of me, there could be ten people spreading false information. Some of the most popular political creators in Canada seemed to be working full-time churning out toxic, misleading content. But no matter which political bubble your feed takes you into, the trust that is built on these platforms is significant, and those of us concerned about the health of our institutions and democracy ignore these platforms at our peril.

Goal17 is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

The Canadian election has had all kinds of strange accusations thrown around, including many calling Mark Carney a socialist. This is strange, indeed, given that he is an alum of Goldman Sachs. But with driverless semi trucks hitting the roads in Texas and autonomous robots building, we might want to consider some of our options.

Goal17 is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

There has been something bothering me for some time as I’ve watched public opinion swing wildly on some longstanding issues, but until the results for Canada’s election came in last night it has felt difficult to put my finger on it.

Now, my area of specialization is in decision making, and as I have focused more and more on decision making around critical societal issues, public opinion has become a critical component of what decisions are possible and how they can be made. But a few years ago, I started seeing some dynamics emerging that we hadn’t been accounting for in our work, and it was only by working in vastly different domains that I was able to better understand what was going on.

Goal17 is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

One basic democratic concept in policy work and in politics is that for anything that you might want to do, whether it’s about tax codes, public health, labor rights or foreign policy, it can’t stray too far from what the general public finds acceptable. If you do something that is too unpopular with too many people for too long, you will get voted out.

There is actually an elegant model for this that has become known as the Overton Window, which describes the range of play that leaders might have on any given issue.

In a given society, at a given moment, there is a range of policies politically acceptable to the mainstream.

A key part of the concept is the role that different players in the system play both in responding to issues that are within the window, but also how they work to shift the window to reflect new perspectives.

Generally, the theory went, politicians will only propose ideas that fall within the window. It falls to think tanks (and others) to propose unpopular things outside of the window in the hope of shifting the window and making the previously unthinkable achievable.

There is an important framing in this model that I really like, and that is the spectrum of acceptability that it lays out, that they intentionally laid out as a vertical so that it didn’t map to the simplistic left/right binary we often use. The spectrum centres on “Policy”, which is something that is so normalized that it can be comfortably enshrined as policy by government, but then has degrees of acceptability that range from there until they fall outside of the window: policy, popular, sensible, acceptable, radical, unthinkable.

The idea goes that when we look at public discourse and public opinion, there will be a window within which a politician can play that can be broad or narrow, but that trying to make the unthinkable into policy won’t be workable.

The Internet Enters the Chat

There are two major factors that have changed the calculations around the Overton Window, in my opinion. The first is the collapse of any shared reality or mainstream, and the second is the effects of personalized media and algorithmic editorialization.

In Musa al-Gharbi’s fascinating book “We Have Never Been Woke”, he outlines a dynamic in which a swing towards a new set of norms in media towards progressive themes creates a response by those that feel left out of those themes to create an alternative information infrastructure. Basically, as media becomes more progressive, people on the political right begin to set up a parallel media environment. This is where we are now with the contemporary right wing and left wing media.

The result of this, however, is that over time, you no longer have a singular “mainstream” like we had in the 90’s, you have increasingly separate media universes with their own parallel realities.

While this is challenging enough, we are also no longer in a network broadcast world, but in a fragmented media landscape where our information diet has radically changed. Whereas your media diet in the 90’s might have been one or two large meals a day - a newspaper and the evening network news - the modern media diet has largely done away with meals and involves constant snacking. And the snacks aren’t necessarily healthy.

Where a proper meal might take a lot of preparation and attention to nutritional balance, the switch to media snacking often consists of highly addictive and over-processed content.

With algorithms and social media platforms, the delivery of this content can also be targeted and tailored to be more addictive to the individual.

Your Own, Personal Window

This fragmentation of the media landscape and the personalization of the media environment means that the idea of the Overton Window is now working with completely different dynamics.

Whereas in Overton’s time, there was a shared, collective conversation in society, where persuasion happened in something like “the public square”, the conversations now happen within increasingly small and tailored bubbles specific to the journey of the individual. Instead of the Window framing what is acceptable to society, it can focus on what is acceptable to you. With every post you like on Facebook, or every explainer video you watch on TikTok or YouTube, your media ecosystem reshapes itself around you accordingly.

This means that rather than public discourse that is being shaped in the political sphere, it is individual thought that is being influenced.

From Discourse to Radicalization and Polarization

Democracy, as countries like Canada practice it, involves periodically asking the population to make a choice of leaders, who will then hold office and push policy for several years before they have to face the electorate again. This means that the stakes are incredibly high for those seeking power to influence that choice at that moment. And because the medium for information is no longer constrained by borders or any traditional institutions, it means that the players in the space can range from individual activists to foreign intelligence agencies, lobbyists, hate groups and leaders of industry.

The idea of “red-pilling” - which uses the famous red pill/blue pill scene from the Matrix as a metaphor - represents the process in social media when acceptance of one piece of an online narrative leads the individual down a rabbit hole of loosely related theories that all “connect” through some shadowy conspiracy. This has gone from a fringe phenomenon to something very widespread. This dynamic is fuelled by recommendation algorithms, as well as the development of intentional radicalization strategies by those pushing the narratives.

An example of this was the “#SaveTheChildren” hashtag, which coopted activism by a real NGO to promote conspiracy on child sex trafficking rings. The strategy was to take an issue that any “normal” person would be concerned about, and to connect it with something more sinister. This was a classic red pill tactic. Once the unwitting victim took the bait on the first part, they would be exposed to progressively more extreme content.

The result is that the spectrum in the Overton Window becomes more like a triangulation challenge: how do you make the popular seem radical, and the unthinkable seem sensible?

Whereas traditional discourse and debate might be more driven by a set of values and principles, the strategy for the unprincipled pursuit of power is to decouple values and principles from individual issues and policies by transforming the mundane into the maniacal. Life-saving vaccines become “forced medical experiments”; walkable neighbourhoods in 15-minute cities become “walled districts in the Hunger Games”. The goal of these is not to propose an alternative policy, but to tie anyone supporting these formerly “mainstream” concepts to sinister networks intent on implementing the unthinkable.

For foreign governments and geopolitical adversaries, this has become a cheap and easy way of seeding deeply polarizing conflict and paralyzing disagreement into their democratic rivals. For politicians attempting to enact policies that would be otherwise totally unpalatable to the public - like slashing social services and funnelling wealth to the rich at the expense of the working class - it is the perfect tactic for building a political movement without having to declare your intentions: you might not vote for me if you knew what I was planning to do with your pension, but you will definitely vote for me if I promise to protect you from my opponent, who feeds on the blood of trafficked children and plans on imprisoning you in your neighbourhood so they can conduct forced medical experiments on you and your loved ones.

Short Term Gain

What I worry about beyond the short-term effectiveness of this strategy is that it is, in the long run, completely corrosive to, and incompatible with, a functioning democracy. Though many Americans recoiled at the violent insurrection on January 6th in Washington that tried to overturn their election, the uncomfortable part is that many of the participants in the insurrection were convinced that they were actually fighting against a very real threat to democracy. The path that we are on right now is one in which that absolute certainty in parallel realities will only become more common and more pronounced.

As we have seen in countries where the West has tried to impose democracy on populations with deep sectarian divisions is that democracy cannot function when different factions view each other as dangerous, existential threats. It is only through a concerted effort to rebuild some semblance of shared truth and reality that we can hope to stay as a functioning democratic society.

In the future, you will own nothing, and you will be happy.

I can’t remember when I first heard this line in the lead up to Canada’s election, but before long, I kept hearing it on repeat as proof that Liberal candidate Mark Carney was part of a shadowy globalist cabal intent on bringing tyranny to Canada.

After hearing it enough times, I felt the need to understand where this was even coming from, as it seemed like an unlikely quote from an ex-Goldman-Sachs/ex-central banker.

Goal17 is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

The source, in case you haven’t heard this one, was a foresight essay written by Danish MP Ida Auken that was published on the website of the World Economic Forum. The post has been taken down by the Forum by now, likely because of the odd controversy surrounding it, but you can find an archive of it here. The point of the essay, which is really more of a fictional vignette, was to take the vague concept of “the sharing economy” and to explore what it would look like if the concept were to be expanded to its fullest extent.

Now, to my knowledge, Carney and Auken don’t even know each other (though maybe they do, who knows). The implication for those who were spreading this conspiracy, however, was that this extreme version of the sharing economy in which nobody owns anything is the official position and secret intention of the World Economic Forum (though I’m not sure why they would publish secret plans on their website) and, because the WEF “controls” leaders around the world, and Carney has attended WEF events, that this essay represents his secret plan for Canada.

This conspiracy has gotten enough traction, clearly, that the WEF has finally just pulled the essay down from their website, which is a shame, because it is rather well done as a thought experiment.

Now as we enter the final stretch of the election campaign, another piece of foresight work is making the rounds, this time, from Policy Horizons Canada. Policy Horizons is the home of foresight in the Canadian Federal government, and the piece in question is called “Future Lives: Social Mobility in Question”.

As the published piece states, in a manner that will be familiar to anyone who has done foresight work:

The scenario below paints a picture of Canada in 2040 in which most Canadians find themselves stuck in the socioeconomic conditions of their birth and many face the very real possibility of downward social mobility.

Now, importantly, as though it needs other be said, the report clearly states that:

While this is neither the desired nor the preferred future, Policy Horizons’ strategic foresight suggests it is plausible. Thinking about future scenarios helps decision-makers understand some of the forces already influencing their policy environment. It can also help them test the future readiness of assumptions built into today’s policies and programs. Finally, it helps identify opportunities to take decisions today that may benefit Canada in the future.

The Problem

The Conservative candidate, Pierre Poilievre, has quoted this paper as a clear projection of a terrifying and dystopian future, made all the more damning because it is “predicted” by the government itself. The paper is now making the rounds on social media as proof that the incumbent Liberal party is intent on the economic enslavement and impoverishment of the Canadian people.

What has happened in both cases is a deliberate misinterpretation of the purpose of the papers in order to provide evidence of a conspiracy. When weaponized in this way, the very elements that give good foresight work its power become a liability.

Forecasting vs. Foresight vs. Policy Direction

Though it shouldn’t really need to be said, these are very different types of work, with a very different intent. Forecasting is the projection of quantitative data into the future to make predictions, like weather forecasting, election forecasting and economic forecasting. It is a discipline of data analysis that identifies trends in past and present data to understand where things are going in the future. It is the business of identifying probable or likely futures.

Foresight might draw from forecasts and quantitative trends, but also has a much more qualitative, speculative flavour to it. It uses scenarios, or narratives, to flesh out what possible futures might look like under certain circumstances. Foresight is a critical component of strategy and decision making processes, because it forces decision makers to consider the full consequences of present decisions and trends when extended into the future, and is often used to force consideration of future possibilities. So while forecasting is all about probable futures, foresight is more about possible futures. While some might be wildly speculative, foresight groups like Policy Horizons a likely to skew more towards plausible future scenarios.

Now, importantly, neither forecasting nor foresight are meant to present what is desirable. That is to say, neither is in the business of making a recommendation. Their purpose, in a decision making process, is to flesh out all the aspects and implications of a possible future so that those making decisions can evaluate whether that future is desirable. If it isn’t, they can then craft a strategy for how to avoid that future. If it is desirable, they can make decisions that they think will make that future outcome more likely.

Policy direction, or policy recommendations, would be outlines of the strategies required to achieve certain outcomes. It’s the work that might come after a forecasting or foresight exercise.

Futures Literacy

Organizations like UNESCO have, for some time, been promoting the idea of “Futures Literacy”, or the idea of improving education around the importance of future considerations in planning and strategy. This would likely be a great addition to high school civics classes, as the spread of conspiracies online using foresight materials suggests a general misunderstanding of what these scenarios are, what they are for, and why they exist.

For my part, foresight is an important part of my practice in designing decision making, especially when the planning environment is as chaotic as it is now. Imagining a set of possibilities for the future, to me, is a critical component of any strategic process, because the context in the future might be very different that the one you are in now. I also use scenarios and foresight to create the space in a decision making process to consider the ethical dimensions of their current decisions when projected into the future or when brought to scale. “If we do this, that is likely to happen. Is that what we want?”

The Chill

What troubles me is that if foresight work is increasingly used as the basis for conspiracy theories, it might put a chill on futures work in decision making while also making organizations less likely to share the foresight work that they have done. I have already had one foresight exercise I have done “leaked” as proof of nefarious play, when in fact I had used it as a way of spurring a conversation on ethics. While the work wasn’t classified, when it was shared, it was presented as if it was, with the scenarios presented as intended outcomes, rather than the ethical dilemmas they were intended to be.

As we’ve seen above, the WEF pulled down Auken’s paper, despite the fact that it is thought-provoking, because the controversy and conspiracy have made it into a distraction.

But I think that given that this trend is being fuelled by a major-party leader in a G7 country should be a wakeup call that critical tools in good decision making and policy making are under attack. Consistent explanations in media about what these reports are, and what they are not, should be the norm, and effective messaging to dissuade leaders from disingenuous references needs to enter the discourse. And futures literacy? I think it’s now more important than ever.

Researchers in Artificial Intelligence often use Chain of Thought (CoT) prompting as a way of getting AI models to improve their reasoning by explicitly laying out the steps taken to answer a specific question. Not only does this focus on the reasoning process help researchers better understand how LLMs arrived at their decision, it also, as it turns out, results in better results.

Like contemporary AI systems, modern governments have reached a point where the decisions being made resemble early LLM responses: it’s impossible to tell what data is being used, the reasoning process is totally unclear and they seem extremely prone to hallucinations.

Goal17 is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

My career, so far, has been focused on decision making; not on telling people what decisions to make, but, rather, on designing the processes by which they make those decisions. Through research and practice, I have seen that through good design and by controlling for a few variables, the quality and speed of decision-making can be vastly improved, especially as our research in psychology and behaviour continue to develop while the quality and quantity of data we can use to inform our decisions has eclipsed what was available to us in the past. There are also vast international communities of practice around a host of methodologies with evidence and practical tools for supporting decision making effectively.

Which is all to say, we have never, in our history, been better equipped than we are today to make excellent decisions that draw from historical experience, weave together the knowledge and perspectives across our many domains of knowledge and are informed by a richness of data beyond anything imaginable even a generation ago.

You would expect that, given all we know and have access to now, we would be in some kind of golden age, plotting a path into the future together with nothing but the laws of physics to constrain our progress.

That, clearly, is not the situation we find ourselves in.

One of the principles we use in our practice is the idea of “designing backwards from desired outcomes”. The idea is to start with the end in mind, then imagine the sequence of conversations and areas of inquiry what would result in the outcome you’re hoping for.

I’m also reminded of another quote by Stafford Beer, which is “the purpose of a system is what it does.” I like how agnostic this statement is; it’s a great analytical prompt because it asks us to ignore what we think a system is for, and to evaluate it based on the actual outcomes, not the intent.

If I were to evaluate our political and decision-making systems using Beer’s lens, our current system’s “purpose” might be characterized by optimizing for short-term decisions that emphasize sentiment over evidence and conflict over consensus.

If I were to approach from the angle of desired outcomes, however, I would imagine that we would like a system that could craft well-informed policies that work in a unified way towards a shared vision for the country and future generations.

There is a considerable gap between these two realities, and I believe that design can play a crucial role in creating the conditions for these kinds of outcomes. But the first step in dealing with a problem is being able to name it.

I believe that the ways in which we approach decision-making in Western democracies no longer represents the best of what we know about structuring decisions, and if we want to enjoy a better future, we need to improve how we make decisions.

We need to improve our infrastructure for establishing shared truths.

We need to build processes for arriving at shared priorities.

We need to establish new norms for accountability in political speech.

We need mechanisms for inter-agency collaboration on complex issues that don’t fall neatly within one category.

We need to develop plans and policies that are responsive to evidence and respectful of future generations.

We need to adapt how we engage in democratic discourse in a way that is open, but can manage threats of foreign interference and misinformation.

We need policy decisions to have the same burden of due process that we expect of verdicts in the justice system.

We need to do all of this in ways that build on fundamental democratic principles.

We already have the tools, techniques and practices to design all of these processes. What we need is the will to put in the work and the courage to experiment. This, to me, is the essence of Process Activism: the knowledge that the decisions we make are shaped by how we make them, and only by designing a better system can we expect to have better results.

Integrating artificial intelligence into your workflow requires you to not only evaluate where and how it can add value, but also, what ethical considerations arise with each new implementation and the policies you might need to put in place as you go. So far I have found that AI can offer a lot of value in collaborative processes, but there are a number of areas where it is easy to violate trust in ways that will harm adoption in the future.

In this post, I wanted to document some of the considerations that have come up so far, and the beginnings of a framework for approaching your own policies. I’ve boiled it down to a few design principles around five key areas: risk, power, privacy, ownership and value.

Goal17 is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

Risk

Principle: Learn with low stakes

There are a lot of risks, both real and imagined, in adopting AI, but they can be difficult to identify until you start using it. There could be security exposure, legal issues, hallucinations or reputational damage, but sometimes those risks can be difficult to fully understand in advance. Each platform has its own risks, but so does every use case. Experimenting and prototyping with low risk data, low risk stakeholders in a low risk environment can help you shape a better understanding of what the possibilities are, and also a more realistic picture of the risks. Trying to design a perfect implementation up front makes it very difficult to understand the full picture; I’ve been playing with hardware, software, process and context in “safe” projects in order to get a better understanding of the pitfalls.

I now gravitate to hardware with failsafes, clear articulations of data retention policies, AI platforms with clears Terms of Service and processing of inputs rather than outputs as a result of this approach.

Power

Principle: Level, don’t amplify, power imbalances

It’s easy to imagine uses for AI that allow you to centralize a lot of control and to optimize, automate or monitor a wider and wider range of inputs. I believe, however, that the AI use cases that will get the most traction are the ones that rebalance power, as opposed to exacerbating existing imbalances. Imagine a call centre, for example. One approach that leans on an existing imbalance would be to deploy chatbots and voice agents that allow the call centre to have fewer staff, and can triage callers before they speak to an agent. If the optimization is purely for the benefit of the company, it will most likely result in even more frustration from callers. An approach that addresses the imbalance would be to have an AI that works as an agent on behalf of the caller, to minimize their time and to negotiate a solution before reaching back out to them.

In a collaborative process, AI can be used to provide more channels for more input and engagement from more people in a meaningful way. Use it to increase, not replace, engagement.

Privacy

Principle: Respect autonomy, earn trust and don’t be creepy

In the workplace, and especially in collaborative settings, it is now possible to process so many inputs that it is very easy to move from “capture” to “surveillance”. I believe that over time, processes that don’t respect the rights of the people who participate will struggle to get buy in, and even those that do will be biased by the behaviour of individuals that know they are being surveilled. Once trust is lost, it is very difficult to get it back, and, further, when this technology is being used in environments where there are low levels of trust to begin with, extra steps will have to be made to get buy in. If you are planning an approach that takes away the autonomy of users or spies on them in a way that wouldn’t otherwise be socially acceptable (would you do this to your family? Friends?) it is very likely to backfire.

While I now use microphones in breakout sessions, for example, I am crafting a clear privacy policy around retention and use of any recordings, and am iterating the system to have no human-in-the-loop so that comments are not traceable to individuals (Chatham House Rule).

Ownership

Principle: Ownership of inputs should correlate to ownership of outputs

Aggregating data to build a new value proposition can lead to the same issues that AI companies have been facing with copyright holders: they are selling the outputs of a model that was created using other people’s inputs. If you are planning on aggregating data, or profiting from the output of aggregation, you should do this in collaboration with those who create the inputs. This is not only the right thing to do, this is an evolving area of law, so protects you from unforeseen exposure in the future.

Value

Principle: Build generative, not extractive, value propositions

AI can be used to extract benefit from others, or it can be used to generate value for everyone involved. While extracting value might be profitable, I think longer term value is to be had with generative value propositions. In a collaborative setting, you might use AI to generate interaction data over time that has value, or build “lock-in” with groups because you hold their data, but I think this will be met with more and more resistance as people become more savvy with the technology. Using AI to build supports for groups that can speed their work and enrich their experience, I think, will get a lot more adoption over time.

In Conclusion

If there was a final principle I would use, it would be this: be willing to show your work. I think that transparency is the best test across the entire workflow. If you’re not comfortable sharing who benefits and how, what technology you’re using, how you’re managing the data, what the data can be used for and how you’re thinking about all the stakeholders in the process, then that should be a gut check that you should make some changes in your approach.

The debate in many Western democracies on what to do about TikTok hinged around a few, very real concerns. The first concern related to the danger of a foreign adversary being able to use very powerful algorithms to shape the discourse around any issue and amplify misinformation. The second concern was in having a foreign adversary gaining deep access to data profiles of so many citizens, as the algorithm is able to generate incredibly rich profiles of its users.

The root problem is one of trust. The debate circled around our growing awareness on how much power and influence modern technologies have over our work, our knowledge, our opinions and, by extension, our democracy and its institutions.

The concern was also not theoretical. We know what China could do with the platform because it is what we are already doing with western tech platforms - US Intelligence and the 5 Eyes currently have the capabilities we are talking about, but because we “trust” each other, those are compromises we live with. The worry is; what happens when an adversary controls key parts of your information infrastructure?

Goal17 is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

I think we will look back at the current moment as the time that American technology dominance began its decline.

The moment that the US made its allies question their trustworthiness, the entire question of collaboration on technology platforms changed. American tech companies have benefited handsomely from global markets, with Microsoft, Google, Amazon, Meta and Apple providing technology used by individuals, governments and businesses around the world. Just as many NATO members grew lazy in investing in their militaries because American protection was already so much easier to rely on, not many countries bothered to create homegrown tech like cloud storage, AI processing, email or productivity software. After all, why would you? American tech is so great, and, how on earth could you compete with Microsoft and Google on core productivity tools?

As a Canadian, I was forced to ask myself, if the US decided to follow through on its threat to annex my country, what might the first step be in that process? Our entire government runs on Azure and Google Cloud - if the US decided to hit the kill switch, our country couldn’t even run payroll, let alone send out any communications.

This should be a wake up call for anyone outside of the US that uses American technology. I’m writing this on a Mac (US), publishing it to Substack (US), my company runs on Google Workplace (US), my AI tools are Anthropic (US), OpenAI (US) and Gemini (US) and my backup cloud storage is Dropbox (US) and Box (US).

Europe is already waking up to their dependence on American technology, with Poland questioning whether Starlink can be trusted as a connectivity partner in Ukraine, sending Eutelsat shares soaring. The threat of Musk deciding to switch off access at key moments was enough to start some serious soul searching, but I think set off some deeper questioning on what it means for unpredictable, untrustworthy parties owning your data access.

With all of the recent patriotic fervour in Canada, I hope that begins to translate into the tech scene, as this country needs to take a long, hard look at what it means to be entirely dependent on a country that is declaring itself as an adversary for all of our information technology. We need a coordinated response, both from our tech industry - where there is a ton of talent - along with industry and government. We need our own DARPA.

For those making decisions for their organizations on tech, it is worth running some scenarios on what the impacts would be of losing connectivity to all of your cloud services, as a very minimum consideration. Build a contingency plan in the short term, but in the long term, we need to build new infrastructure.

The tech market is never going to be the same.