Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: How might we align transformative AI if it’s developed very soon?, published by HoldenKarnofsky on August 29, 2022 on The AI Alignment Forum.

This post is part of my AI strategy nearcasting series: trying to answer key strategic questions about transformative AI, under the assumption that key events will happen very soon, and/or in a world that is otherwise very similar to today's.

This post gives my understanding of what the set of available strategies for aligning transformative AI would be if it were developed very soon, and why they might or might not work. It is heavily based on conversations with Paul Christiano, Ajeya Cotra and Carl Shulman, and its background assumptions correspond to the arguments Ajeya makes in this piece (abbreviated as “Takeover Analysis”).

I premise this piece on a nearcast in which a major AI company (“Magma,” following Ajeya’s terminology) has good reason to think that it can develop transformative AI very soon (within a year), using what Ajeya calls “human feedback on diverse tasks” (HFDT) - and has some time (more than 6 months, but less than 2 years) to set up special measures to reduce the risks of misaligned AI before there’s much chance of someone else deploying transformative AI.

I will discuss:

Why I think there is a major risk of misaligned AI in this nearcast (this will just be a brief recap of Takeover Analysis).

Magma’s predicament: navigating the risk of deploying misaligned AI itself, while also contending with the risk of other, less cautious actors doing so.

Magma’s goals that advanced AI systems might be able to help with - for example, (a) using aligned AI systems to conduct research on how to safely develop still-more-powerful AI; (b) using aligned AI systems to help third parties (e.g., multilateral cooperation bodies and governments) detect and defend against unaligned AI systems deployed by less cautious actors.

The intended properties that Magma will be seeking from its AI systems - such as honesty and corrigibility - in order to ensure they can safely help with these goals.

Some key facets of AI alignment that Magma needs to attend to, along with thoughts about how it can deal with them:

Accurate reinforcement: training AI systems to perform useful tasks while being honest, corrigible, etc. - and avoiding the risk (discussed in Takeover Analysis) that they are unwittingly rewarding AIs for deceiving and manipulating human judges. I’ll list several techniques Magma might use for this.

Out-of-distribution robustness: taking special measures (such as adversarial training) to ensure that AI systems will still have intended properties - or at least, will not fail catastrophically - even if they encounter situations very different from what they are being trained on.

Preventing exploits (hacking, manipulation, etc.) Even while trying to ensure aligned AI, Magma should also - with AI systems’ help if possible - be actively seeking out and trying to fix vulnerabilities in its setup that could provide opportunities for any misaligned AI to escape Magma’s control. Vulnerabilities could include security holes (which AI systems could exploit via hacking), as well as opportunities for AIs to manipulate humans. Doing this could (a) reduce the damage done if some of its AI systems are misaligned; (b) avoid making the problem worse via positive reinforcement for unintended behaviors.

Testing and threat assessment: Magma should be constantly working to form a picture of whether its alignment attempts are working. If there is a major threat of misalignment despite its measures (or if there would be for other labs taking fewer measures), Magma should get evidence for this and use it to make the case for slowing AI development across the board.

Some key tools that could help Magma with all of the above:

Decoding and manipulating in...

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: [Cause Exploration Prizes] Preventing stillbirths, published by Denise Melchin on August 29, 2022 on The Effective Altruism Forum.

This essay was submitted to Open Philanthropy's Cause Exploration Prizes contest.

If you're seeing this in summer 2022, we'll be posting many submissions in a short period. If you want to stop seeing them so often, apply a filter for the appropriate tag!

Author's note: This cause area exploration is far from the depth and clear conclusions I would want and is in parts clearly in draft mode. Open Phil encouraged me to submit regardless and publish properly later. I will do this to encourage myself to actually finish it at a later date and not forget the post in the drawer.

This submission might change focus later: I could imagine improved antenatal and labour care to have wider benefits on child health than reducing stillbirths.

Summary

Two million babies are stillborn every year. A baby is called stillborn if they die after 28 weeks of pregnancy. Most charitable interventions focus on children already born despite babies being much more vulnerable in the last trimester. Almost half of all stillborn babies were still alive until labour started, passing away hours or mere minutes before they were born. Stillbirths are usually not given a DALY weighting so are ignored in cost-effectiveness analysis. Interventions reducing maternal and child deaths often reduce stillbirths as well, so they are more cost-effective than they first appear.

Intragovernmental bodies as well as philanthropists have shown much less interest in reducing stillbirths compared to maternal and child deaths. This gives philanthropists the opportunity to save more lives by focusing on stillbirths.

To a first approximation, children become less vulnerable as they get older. Adjusting for the length of the time period, the stillbirth rate is much higher than the infant mortality rate which in turn is much higher than the under 5’s mortality rate. Although birth is the riskiest period.

As philanthropists we have already noticed that focusing on young children is often most cost-effective. All recommended GiveWell charities above the GiveDirectly bar are focusing on the lives of young children. Trying to reduce stillbirths by improving care of pregnant women would just be going one step further in this direction.

Uncertainties

Major

This submission primarily discusses stillbirths. But the potential interventions to reduce stillbirths have wider benefits. I do not know what fraction of the benefits of the interventions actually accrue as stillbirth prevention. Therefore it is possible a later publication will pivot into antenatal interventions and/or labour care more widely.

As of now, his post is missing plenty of potential interventions, instead just giving a cursory overview, as well as lacking cost-effectiveness estimates. Therefore it is hard to say whether it is a plausible competitive cause area candidate.

Minor

Some data is a bit outdated. This seems particularly risky due to geographic shift in stillbirths.

I have not paid enough attention to different risk factors by geography.

Importance

Two million babies are stillborn every year. A baby is called stillborn if they die after the 28th week of pregnancy but before they are born. If born in a rich country, most of these babies would live. In the UK, a baby born at 28 weeks is around 90% likely to survive up to virtually a hundred percent if full-term.

This is especially true for the almost half of all stillbirths in which the baby only dies after labour has already started.

A sharp discontinuity in health cost-effectiveness analysis such that a baby passing away minutes after birth should be given the full QALY weighting while a baby passing away minutes before should be ignored seems hard to defend.Babies passing away...

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: An Audio Introduction to Nick Bostrom, published by peterhartree on August 29, 2022 on The Effective Altruism Forum.

Team Radio Bostrom has just released An Audio Introduction to Nick Bostrom.

It offers a deep dive into Nick Bostrom's ideas, including: existential risk, wise philanthropy, the ethics of AI, transhumanism, and the case for speeding up some investments in technology while slowing down others.

We'd love to hear your thoughts!

How you can help

If you'd like to help spread the world, please like and retweet this:

And please, share your thoughts on how we could improve this. Or tell us some other people who might be good for a similar series...

See also: Radio Bostrom launch announcement.

Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Should GiveWell still recommend GiveDirectly?, published by Shakeel Hashim on August 29, 2022 on The Effective Altruism Forum.

This is a couple weeks old, but I don't think it's been shared here yet. For context: GiveWell recently took GiveDirectly (which runs cash-transfer programs) off its "top charities" list. GiveDirectly wrote a very gracious and interesting response. One section particularly jumped out:

GiveWell is a charity evaluator with the stated goal of “finding and recommending a small number of outstanding giving opportunities to help donors save or improve lives the most with their gifts.” As one of the largest private funders in global health and development, they’ve made a huge difference in the lives of millions and been a positive influence in the sector. Their priority is to maximize the impact of the funds they direct from donors (~$500M in 2021), so they choose cost-effectiveness cutoffs based on the amount they expect to move.

But there’s a lot more money at stake. Official development assistance is $178B a year, and U.S. charitable giving alone is another $484B a year. Recently, Mackenzie Scott gave away $4B in 9 months, and Elon Musk debated how to use $6B with the head of the World Food Programme. We want an approach to giving well that has answers for budgets at these scales.

GiveWell would say they’re focused on prudently allocating the money they expect to direct, and if they received even more funds, they would figure out how to allocate those well too. We expect that they would. But GiveWell is not just any other donor — it is the premier, trusted voice on how to give. Structuring GiveWell’s recommendations only around the funds they expect to direct, at best, says nothing about the vast majority of funds that could help people living in extreme poverty. At worst, it suggests these funds can’t do much good.

This strikes me as a good point. It makes complete sense to direct the funds we have, right now, to charities that are 10x as cost-effective as cash transfers. But my (very uncertain!) understanding is that those programs will run out of room for funding at some point. So it does seem weird for EA organisations to not say anywhere that if we had billions more in resources we could also massively scale up cash transfers. That seems particularly worth saying because by highlighting that, we might be able to encourage those billions to come our way. To take GiveDirectly's example, we could say to Mackenzie Scott: if you give us $4B tomorrow to spend on global health, we can give all of it away really effectively through a mix of malaria nets, malaria chemoprevention, vitamin A supplements, conditional cash transfers and unconditional cash transfers.

But, like I say, I'm uncertain here, and would love other people's thoughts. One idea: perhaps GiveWell should have a "If we had X amount of money we'd do this" page, with milestone targets?

Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: (My understanding of) What Everyone in Technical Alignment is Doing and Why, published by Thomas Larsen on August 29, 2022 on The AI Alignment Forum.

Epistemic Status: My best guess

Epistemic Effort: ~50 hours of work put into this document

Contributions: Thomas wrote ~85% of this, Eli wrote ~15% and helped edit + structure it. Unless specified otherwise, writing in the first person is by Thomas and so are the opinions. Thanks to Miranda Zhang, Caleb Parikh, and Akash Wasil for comments. Thanks to many others for relevant conversations.

Introduction

Despite a clear need for it, a good source explaining who is doing what and why in technical AI alignment doesn't exist. This is our attempt to produce such a resource. We expect to be inaccurate in some ways, but it seems great to get out there and let Cunningham’s Law do its thing.

The main body contains our understanding of what everyone is doing in technical alignment and why, as well as at least one of our opinions on each approach. We include supplements visualizing differences between approaches and Thomas’s big picture view on alignment.

The opinions written are Thomas and Eli’s independent impressions, many of which have low resilience. Our all-things-considered views are significantly more uncertain.

A summary of our understanding of each approach:

Problem FocusCurrent Approach SummaryModel splinteringSolve extrapolation problems. Inaccessible informationELK + LLM power-seeking evaluationLack of good interpretability tools (?)Interpretability + HHH + augmenting alignment research with LLMsBrain-like AGI SafetyUse brains as a model for how AGI will be developed, think about alignment in this contextEngaging the ML community, many technical problems Technical research, Infrastructure, and ML community field-building for safetyOuter alignment, though CHAI is diverseImprove CIRL + many other independent approaches. Suffering risksFoundational game theory researchInner alignmentInterpretability + automating alignment research with LLMsMany including scalable oversight and goal misgeneralizationMany including Debate, ERO, and discovering agents.

Multipolar failure from lack of coordinationVideo gameDeceptionGet the reasoning of the AGI to happen in natural language, then oversee that reasoningMany (?)Incubate new, scalable alignment research agendasMany including deception, the sharp left turn, corrigibility is anti-naturalMathematical research to resolve fundamental confusion about the nature of goals/agency/optimizationScalable oversightRLHF / Recursive Reward Modeling, then automate alignment researchScalable oversightSupervise process rather than outcomes + augment alignment researchersInner alignment (?)Interpretability + Adversarial Training Being able to robustly point at objects in the worldSelection Theorems based on natural abstractionsInstilling inner values from an outer training loopFind patterns of values given by current RL setups and humans, then create quantitative rules to do thisDeceptionCreate standards and datasets to evaluate model truthfulness

Approach

Aligned AI

ARC

Anthropic

Brain-like-AGI Safety

CAIS

CHAI

CLR

Conjecture

DeepMind

Encultured

Externalized Reasoning Oversight

FAR

MIRI

OpenAI

Ought

Redwood

Selection Theorems

Team Shard

Truthful AI

Previous related overviews include:

Neel Nanda's My Overview of the AI Alignment Landscape

Evan Hubinger's An overview of 11 proposals for building safe advanced AI

Larks' yearly Alignment Literature Review and Charity Comparison

Nate Soares' On how various plans miss the hard bits of the alignment challenge

Andrew Critch's Some AI research areas and their relevance to existential safety

80,000 Hours’ list of organizations working in the area

Aligned AI / Stuart Armstrong

One of the key problems in AI safety is that there are many ways for an AI to gener...

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Robert Long on Why You Might Want To Care About Artificial Sentience, published by Michaël Trazzi on August 28, 2022 on The Effective Altruism Forum.

I talked to Robert Long, research fellow at the Future of Humanity Institute, working at the intersection of the philosophy of AI Safety and consciousness of AI. Robert has done his PhD at NYU, advised by David Chalmers.

We talk about the recent LaMDA controversy about the sentience of large language models (see Robert's summary), the metaphysics and philosophy of consciousness, artificial sentience, and how a future filled with digital minds could get really weird.

Below are some highlighted quotes from our conversation (available on Youtube, Spotify, Google Podcast, Apple Podcast). For the full context for each of these quotes, you can find the accompanying transcript.

Why Artificial Sentience Is A Pressing Issue

Things May Get Really Weird In The Near Future

“Things could get just very weird as people interact more with very charismatic AI systems that, whether or not they are sentient, will give the very strong impression to people that they are. I think some evidence that we will have a lot of people concerned about this is maybe just the fact that Blake Lemoine happened. He wasn’t interacting with the world’s most charismatic AI system. And because of the scaling hypothesis, these things are only going to get better and better at conversation.”

“If scale is all you need, I think it’s going to be a very weird decade. And one way it’s going to be weird, I think, is going to be a lot more confusion and interest and dynamics around AI sentience and the perceptions of AI sentience.”

Why illusionists about consciousness still have to answer hard questions about AI welfare

“One reason I wrote that post is just to say okay, well here’s what a version of the question is. And I’d also like to encourage people, including listeners to this podcast, if they get off board with any of those assumptions, then ask, okay, what are the questions we would have to answer about this? If you think AI couldn’t possibly be conscious, definitely come up with really good reasons for thinking that, because that would be very important. And also would be very bad to be wrong about that.

If you think consciousness doesn’t exist, then you presumably still think that desires exist or pain exists. So even though you’re an illusionist, let’s come up with a theory of what those things look like.”

On The Asymmetry of Pain & Pleasure

“One thing is that pain and pleasure seem to be in some sense, asymmetrical. Its not really just that', it doesn't actually seem that you can say all of the same things about pain as you can say about pleasure, but just kind of reversed. Like pain, at least in creatures like us, seems to be able to be a lot more intense than pleasure, a lot more easily at least. It's just much easier to hurt very badly than it is to feel extremely intense pleasure.

And pain also seems to capture our attention a lot more strongly than pleasure does, like pain has this quality of you have to pay attention to this right now that it seems harder for pleasure to have. So it might be to explain pain and pleasure we need to explain a lot more complicated things about motivation and attention and things like that.”

The Sign Switching Argument

"One thing that Brian Tomasik has talked about and I think he got this from someone else, but you could call it the sign switching argument. Which is that you can train RL agent with positive rewards and then zero for when it messes up or shift things down and train it down with negative rewards. You can train things in exactly the same way while shifting around the sign of the reward signal. And if you imagined an agent that flinches, or it says "ouch" or things like that, it'd be kind of weird if you were...

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: The Expanding Moral Cinematic Universe, published by Raemon on August 28, 2022 on LessWrong.

A subtheme for me this year has been grappling with how to reconcile different facets of my morality.

Part of this has to do with reconcile the virtues of "Protect yourself, maintain slack, be aligned with yourself and your community" and "But, maybe the world is metaphorically and/or literally on fire. How do you want to relate to that?"

Part of this has to do with "man, the story of the Expanding Circle of Concern may not be as nice as I thought, in ways that fundamentally challenge my conception of what 'my particular morality' even means, or whether it is coherent." (This has gone hand-in-hand with me realizing my conception of 'community' may also not really be coherent, which has been challenging for my identity).

Along the way, I've watched some movies and shows that feel like they grapple with a particulate facet of the expanding circle of concern, and they're accumulating into a private (public now) headcanon of the Expanding Moral Cinematic Universe. Narratives that feature characters in a harsh, bloody world who inch their little corner of the universe forward as a place where friendship and cooperation can form. Less a sea of blood and violence and mindless replication.

Here are three vignettes, written over the course of the last year (originally on Facebook, later posted on shortform).

Part I: The Fox and the Hound

Originally posted January 24th on Facebook.

I watched Disney's The Fox and The Hound a few weeks several months ago. I cried a bit.

While watching the movie, my girlfriend commented "so... they know that foxes are also predators, right?" and, yes. They do. This is not a movie that was supposed to be about predation except it didn't notice all the ramifications about its lesson. This movie just isn't taking a stand about predation.

This is a movie about... kinda classic de-facto tribal morality. Where you have your family and your tribe and a few specific neighbors/travelers that you welcomed into your home. Those are your people, and the rest of the world... it's not exactly that they aren't people, but, they aren't in your circle of concern. Maybe you eat them sometimes. That's life.

Copper the hound dog's ingroup isn't even very nice to him. His owner, Amos, leaves him out in a crate on a rope. His older dog friend is sort of mean. Amos takes him out on a hunting trip and teaches him how to hunt, conveying his role in life. Copper enthusiastically learns. He's a dog. He's bred to love his owner and be part of the pack no matter what.

My dad once commented that this was a movie that... seemed remarkably realistic about what you can expect from animals. Unlike a lot of other disney movies it didn't require suspending disbelief much.

A baby fox and hound might totally play together because they haven't figured out yet that they're supposed to be enemies.

If that hound then went away for 6 months to learn to hunt, and came back, it might initially be hesitant to hunt down its fox friend, out of a vague/confused memory.

But interspecies friendship isn't that strong, and doing-what-your-species/tribe does is often stronger, and yeah later on the hound is like "okay I guess we're hunting this fox now. It's what master wants. I do what master says, that's who I am."

And then...

...well, and then the hound gets attacked by a bear. And the fox comes back to save him. And on one hand, I bet 99+% of foxes would not do that. But, having seen a bunch of youtubes of animals doing impressive things, I'm willing to buy that the occasional hero foxes exist, and I'm willing to buy a fox who remembers enough of his interspecies friend to intervene bravely.

(weakly held, if someone who knew a lot more about foxes than me was like "nope, this is outside the space of what...

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: What domains do you wish some EA was an expert in?, published by Jack R on August 28, 2022 on The Effective Altruism Forum.

Similarly: what domains do you wish someone would take a deep dive into and write about their learnings?

A few months ago, I was chatting with Richard Ngo when we concluded that perhaps more EAs should learn things that no one else in the EA community is learning. In other words, it would be good if EAs' knowledge were more uncorrelated.

I then asked him one of these questions, and his answer made me consider leaning into curiosity about corporate governance (one of his answers) with the aim of writing a post about my learnings and/or mentioning my findings to him.

And so I figured I'd ask everyone - perhaps someone will look into it for you and give you some answers.

Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: EA on nuclear war and expertise, published by bean on August 28, 2022 on The Effective Altruism Forum.

I believe that there may be systematic issues with how Effective Altruism works when dealing with issues outside the core competence of its general membership, particularly in areas related to defense policy. These areas are complex and unlike many fields EA operates in, are often short of good public-facing explanations, which may make them hard for outside researchers, such as those from EA organizations, to understand in short order. This raises the possibility of getting basic, uncontroversial details wrong, which can both render analyses inaccurate and ruin the credibility of EA with experts in the field.

A bit of background: I've been fascinated by the defense world for over 20 years, and have spent the last 5 working for a major defense contractor. I write a blog, mostly covering naval history, at navalgazing.net, and it was this work which brought me into contact with what appears to be the most in-depth evaluation of nuclear war risks by EA, work conducted by Luisa Rodriguez for Rethink Priorities.

Unfortunately, while this was a credible effort to assess the risks of nuclear war, unfamiliarity with the field meant that a number of errors crept in, ranging from the trivial to the serious. The most obvious are in the analysis of the survivability of the US and Russian nuclear arsenals.

For instance, when discussing the sea-based deterrent, the article states that "[America's] submarines’ surfaces are covered in plastic, which disperses radar signals instead of reflecting them." This is clearly a reference to the anechoic tiles used on modern submarines, but these are intended to protect from sonar, not radar. This probably traces to the source used, an article that was confusingly written by someone who clearly doesn't know all that much about submarine design and ASW sensors, but it is exactly the sort of error which flags an article as being written by someone who doesn't really know what they're talking about.

But that's merely a nitpick, and there's also a very basic flaw with the assumption that nearly all of the warheads aboard SSBNs will survive. While I completely agree that any submarine at sea is nearly invulnerable (and am in fact rather more skeptical than some of the sources that improved technology will render the seas transparent) a substantial fraction of the SSBN force is in port, not at sea. The US attempts to minimize this by providing each submarine with two crews and swapping them out, but even with this, each operational submarine still spend about a third of its time in port between patrols. (It will spend more time in overhauls, but sending ships to the yard with missiles aboard is considered bad form, so those warheads probably won't count against the US total.) How these would fare in a war would depend heavily on the situation leading up to the outbreak of war.

If Russia launched a surprise attack, the bases at Bremerton and Kings Bay would undoubtedly be the highest-priority targets. If there had been substantial warning, then most would likely have been ordered to sea to strengthen the US deterrent.

The description of the strategic bomber force bears little connection to the reality of said force and contains even more jarring errors, such as describing the bombers as "air-based". The second paragraph begins with the following statement: "While many strategic bombers are concentrated at air bases and aircraft carriers, making them potentially convenient to target, early warning systems would likely give US pilots enough time to take off before the nuclear warheads arrived at their targets and detonated."

Every part of this sentence is wrong. While the US Navy did operate aircraft that were at least arguably strategic bombers, the...

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: The History of AI Rights Research, published by Jamie Harris on August 27, 2022 on The Effective Altruism Forum.

Explanation for the Forum

I previously created a formal literature review of research on "The Moral Consideration of Artificial Entities" and a brief, informal primer on "The Importance of Artificial Sentience." Tldr; seeking to expand the moral circle to include artificial sentient beings seems an important, neglected, and potentially tractable method for improving the long-term future and avoiding risks of astronomical suffering.

I wanted to follow up on those projects with a history of relevant research that was not constrained by a formal, systematic methodology; this would allow me to dive into citation trails and really get a sense of how the field has developed.

The report has a few goals:

To contextualize and connect the relevant streams of research in order to encourage further contributions to this growing field, since field building seems like a promising stepping stone towards moral circle expansion.

To better understand which levers can be pulled on to further increase interest in this specific topic. This could inform field-building work of Sentience Institute and other relevant actors.

More generally, to provide a detailed case study of the development of a new research field. I hoped that this might be useful to other efforts to create new research fields inspired by the principles of effective altruism and longtermism, such as AI alignment, global priorities research, and welfare biology.

The full report includes a detailed chronology of contributions. I expect this to be valuable for people seeking to reduce suffering risks via moral circle expansion, or otherwise exploring questions relating to artificial sentience/digital minds.

However, to keep the forum post brief and focus on the most relevant parts for people seeking to encourage the development of other research fields, I've stripped out the "methodology" and "results" sections, plus the reference list and any footnotes. Please refer to the full report if you're interested in those bits!

Thanks to Abby Sarfas, Ali Ladak, Elise Bohan, Jacy Reese Anthis, Thomas Moynihan , and Joshua Gellers for feedback.

Summary

This report documents the history of research on AI rights and other moral consideration of artificial entities. It highlights key intellectual influences on this literature as well as research and academic discussion addressing the topic more directly.

We find that researchers addressing AI rights have often seemed to be unaware of the work of colleagues whose interests overlap with their own.

Academic interest in this topic has grown substantially in recent years; this reflects wider trends in academic research, but it seems that certain influential publications, the gradual, accumulating ubiquity of AI and robotic technology, and relevant news events may all have encouraged increased academic interest in this specific topic.

We suggest four levers that, if pulled on in the future, might increase interest further: the adoption of publication strategies similar to those of the most successful previous contributors; increased engagement with adjacent academic fields and debates; the creation of specialized journals, conferences, and research institutions; and more exploration of legal rights for artificial entities.

Figure 1: Cumulative total of academic publications on the moral consideration of artificial entities, by date of publication (from Harris & Anthis, 2021)

Figure 2: A summary chronology of contributions to academic discussion of AI rights and other moral consideration of artificial entities

Synthesis Moral and social psych Social-relational ethics HCI and HRI Machine ethics and roboethics Floridi’s information ethics Transhumanism, EA, and longtermism Legal rights...