The Source - An AI Deep Dive Podcast

On Episode 17 of “The Source, an AI Deep Dive” podcast, Jane Context and Mike Loop ask who pulls the plug when AI goes rogue. Frontier AI models need two kill switches. The logical one stops the model, revokes access and suspends agents. The physical one cuts power and network at the data center, for when the software can’t be trusted.

Who should hold the switch: the AI lab, the government, or an independent body?

On Episode 16 of “The Source, an AI Deep Dive” podcast, Jane Context and Mike Loop explain how AI agents get hijacked, and why the risk grows with every tool and permission you give them.

Four threats sit behind most incidents: prompt injection, lateral movement, data exfiltration and excessive agency. Each is manageable alone. Chained together, a single hidden instruction in a document the agent reads can reach systems it was never meant to touch.

On Episode 15 of "The Source, an AI Deep Dive" podcast, Jane Context and Mike Loop make the case that AI strategy is cultural before it's technical, and confidence is the metric that proves it's actually working.

Most AI dashboards track speed, scale, and cost savings. None of it means anything if your people don't trust the system.

The real KPI of AI strategy is confidence: that the AI is fair, that it can be challenged, that it's a partner instead of a black box.

If your teams are rubber-stamping outputs without understanding them, you haven't scaled intelligence. You've scaled guesswork.

Confident in your AI, or just compliant?

The Source: https://www.linkedin.com/posts/activity-7414274595333267456-ziRK

On Episode 14 of "The Source - An AI Deep Dive Podcast", Jane Context and Mike Loop walk through the corporate AI soap opera nobody wants to admit they're starring in.

Every company insists they're different. Then they hit the same seven stages anyway: Denial. FOMO. Overinvestment. Panic. A sudden discovery of governance. A breakthrough where humans stay in the loop. And finally, maturity, when AI gets boring and nobody talks about it anymore.

Most companies are stuck somewhere between stage 2 and stage 5, buying platforms nobody owns and discovering that a spreadsheet named "final_v27_really_final.xlsx" isn't a data strategy.

What stage is your company in?

The Source: https://www.linkedin.com/posts/activity-7424421425622646784-t9Bo

On Episode 13 of "The Source - An AI Deep Dive Podcast", Jane Context and Mike Loop ask why companies are cutting the exact training that builds senior engineers.

AI is quietly removing the step that builds engineering judgment.

Skip that step long enough and you don't have a bench. You have a countdown.

The Source: https://www.linkedin.com/posts/pedromartinspt_ai-engineeringleadership-aiadoption-share-7503575975008534529-i2tn/

On Episode 12, Jane Context and Mike Loop dig into The London School of Economics and Political Science (LSE)and Protiviti research showing multigenerational teams augmented by AI aren't just more inclusive. They're 17% more productive than low-diversity teams.

70% of employees use AI at work. Only 49% trust it. That gap isn't generational. Gen Z adoption (82%) beats Baby Boomers (52%), but train anyone properly and 93% start using AI, regardless of age.

The question isn't who's ready for AI. It's who you've actually trained.

The Source: https://www.linkedin.com/posts/pedromartinspt_ai-artificialinteligence-strategicleadership-share-7398881240730476545-oJ4U/

On Episode 11, of "The Source - An AI Deep Dive Podcast", Jane Context and Mike Loop walk through why AI readiness depends on treating these five strategies as one system instead of five arguments.

Data lakes. Data mesh. Data fabric. MDM. Data products.

Most teams treat these as competing buzzwords, or worse, five separate projects fighting for budget.

Picture a city water system instead. A data lake is the reservoir. MDM is the treatment plant, the place that turns raw data into something trustworthy. Data mesh puts local teams in charge of their own pipes. Data fabric connects those teams into one network. Data products bottle the result for whoever needs it.

Skip any one piece and the system breaks somewhere: a reservoir with no treatment plant, pipes that don't connect, bottled water nobody trusts.Which one does your organization skip?

The Source: https://www.linkedin.com/posts/juan-jose-p-0138b7_datastrategy-dataarchitecture-aireadiness-activity-7488990778636177409--mca

Governance built to catch a bad stat or a biased summary you can edit and move on has no answer for an agent that misreads a decimal point and pays out a batch of vendor overpayments before anyone reviews it.

Agents don't give you that chance. Give a system the ability to book travel, edit a record, or move money, and it stops producing drafts. It produces outcomes.

Episode 10 of The Source: Jane Context and Mike Loop on what agent governance actually has to answer: not "did it say something wrong," but what it's allowed to touch, who approved that scope, and what happens after it acts.

The Source: https://www.linkedin.com/feed/update/urn:li:activity:7495434396884443137/

Every undocumented AI pilot is debt you haven't priced yet.

Episode 9 of The Source: Jane Context and Mike Loop break down AI governance debt - pilots with no owner, approvals buried in a Slack thread, a model nobody remembers greenlighting.

It doesn't feel urgent. Not until an auditor, or a customer, asks who approved it.

How many ungoverned AI pilots is your organization running: a real number, or a guess?

The Source: https://www.linkedin.com/feed/update/urn:li:activity:7490345826502819841/

Every baby born this year will grow up without a "before AI."

Episode 8 of The Source is live. Jane Context and Mike Loop look at what happens when intelligence stops being a tool you adopt and becomes the baseline you're born into — as ordinary and invisible as electricity.

For this generation, the hard part won't be learning. It'll be deciding what deserves their trust and attention in a world where information is infinite and certainty is scarce. And the skills that stay human — judgment, taste, empathy, leadership under ambiguity — become the ones that matter most.

We're the last generation that remembers life before the machines. They won't understand how we managed without them.

The Source: https://www.linkedin.com/posts/activity-7455179089835012096-HAgQ

#AI #AILeadership #FutureOfWork #AIAdoption #Podcast