Thanks, I'm looking forward to this! Some questions that seem worth considering to me are:
1. Is AGI likely to lock in values? (if so it's probably bad for animals) 2. Is the answer to this question even knowable? (a lot of what I've heard on the topic has been like "AI could mean X but also not X") 3. If AGI is good/bad, how steerable is it? (e.g. maybe making sure that AGI goes well for humans is actually much easier)
Thanks for organizing this debate week — it looks very valuable.
We’d like to suggest a piece we recently wrote that may be relevant to the discussion: AI Can Help Animal Advocacy More Than It Can Help Industrial Farming. It examines how the structural limits of intensive animal farming may constrain the gains AI can bring to that sector, while AI could be far more transformative for animal welfare through advocacy, transparency, welfare measurement, and alternatives.
While the piece focuses on near-term AI rather than AGI directly, we think it complements the debate well, especially around questions about whether AI might prolong factory farming or instead help expose and reduce animal suffering.
If helpful, we’d be glad for it to be considered for the reading list.
If AGI goes well for humans, it’ll probably [i.e. ≥70%] go well for animals
I think it doesn't really make sense to do a sliding-scale vote on ≥70%. If my credence on (AGI goes well for animals | AGI goes well for humans) is 60%, then I'm just a no; if my credence is 80%, then I'm just a yes.
One way people could interpret the sliding-scale is expressing their confidence/stability in their judgment about whether the probability is greater or less than 70%. But that's somewhat deranged and it's not clear how to make it precise and I think everyone will just be confused.
It would be totally reasonable to vote on a scale from 0% to 100% for P(AGI goes well for animals | AGI goes well for humans), rather than voting from fully-disagree to fully-agree for ≥70%. Obviously that requires a little recoding. But making the voting scale more flexible, rather than just from fully-disagree to fully-agree, will benefit other debates too.
During the week of March 23 – 29[1], we’ll be debating the statement:
“If AGI goes well for humans, it’ll probably[2] go well for animals”
I'm open to minor changes to the phrasing of this statement, but the general shape is locked in, since people might want to start writing posts for the debate week this weekend.
In this post, I’ll give some background on why this topic matters and share some tips and resources to help you form a view on the topic, before the debate week begins. But first:
How does a debate week work?
During an EA Forum debate week, we put a debate slider on the frontpage banner. It looks something like this:
Anyone who logs into the Forum can then vote on this banner. When you vote, you’re also asked if you want to leave a comment, which will appear on a discussion thread, and be attached to your icon on the banner.
Alongside the conversation in the discussion thread, Forum users also write posts which might influence readers towards agreeing or disagreeing with the statement.
A relatively tiny group of people put serious odds on AGI radically transforming the world. Almost nobody thinks about how to influence outcomes for the better in a post-AI world. Even in that group, the conversation is focused on humans, and/or digital minds.
And yet, the consequences of a post-AGI world might be terrible for animals. The field of AI alignment aims to produce an AGI which is aligned to human values, but such values, as they stand, do not necessary include concern for animal suffering.
Even with an aligned AGI then, we could be facing a future full of bad outcomes for animals: continued farming, the spread of wild animal communities to new planets, or perhaps simply a lack of attention to what a flourishing future looks like for species other than human.
It's also absolutely possible that all marginal EA-influenced work should go towards surviving the transition to a post-AGI world/ general flourishing post-AGI (for that discussion, see last debate week). But I'd argue that the sheer number of animals, and the almost total lack of serious thinking[3] about the impact of AGI on them justifies us entertaining an alternative. That's what this week is for.
Questions to consider
To help you form a view on the debate statement, I'll list a few (non-exhaustive) questions which might nudge you one way or the other.
As a reminder, the statement is: “If AGI goes well for humans, it’ll probably[2] go well for animals”
How transformative do I expect AGI to be? Are we talking about a productivity speed-up, or a total reshaping of society?
Many of the concerns listed in this post are more relevant in a less transformed world - for example the risk of marginally more efficient factory farming increasing stocking density and thus suffering.
In worlds where AGI goes well, does that mean a total moratorium until we have a better understanding of AGI, or is it possible to ensure 'goes well' levels of alignment in the next decade?
What does it mean to go well for animals? Does the ideal future contain animals? In what way?
Would an AGI 'aligned to human values' be bad for animals? Are human values bad for animals?
Suggest more questions in the comments, I might add them.
Reading list
A list of posts which might help inform your understanding of the topic, and influence your cruxes.
Lizka and Ben expect a very transformative form of AI which will change the global economy and overturn most of our assumptions. One interesting takeaway: they don’t expect factory farming to exist in this transformed post-AI future, so they don’t think we should align present-day AIs to particular views on farming.
Please suggest more in the comments, I'll add them.
How to contribute to debate week
Consider writing a post! A great debate week post illuminates a consideration which can help the reader change their mind, or deepen their understanding of the central debate topic. A great debate week post does not have to be directly about the statement.
Message me or comment below if you have suggestions for people who should write posts or take part in the discussion during the debate week. I'll be reaching out to a bunch of possible authors next week.
PS- thanks to everyone who has already contributed to this debate week, including all the commenters here, and Forum users who voted on our Forum-Choice debate week poll.
We need a ‘probably’ here because the claim is that in most possible worlds where AGI goes well, it’ll go well for animals. Without 'probably', the claim is far too certainly phrased - no one should actually agree with it. If you'd like a precise definition, let's call 'probably' 70%.
AI Use Note: Main body text entirely human written. Claude (Opus 4.8) helped develop models of animal life histories in the appendix.
Cross-posted from Good Structures.
Executive Summary
* Animal advocates sometimes make claims like “there are X of this animal...
Summary
Back in November 2023 I posted here to launch Spiro and raise our first $198k. Two and a half years later this is an update and a fundraiser for the next step.
The short version: we've now reached over-5,900 people with TB preventive medicine, including over 3,000 children under five years old. Our early results have held up well an...
Adapted from my Substack, Funding Anthropalypse.
Short version: if you want a share of the coming Anthropic and OpenAI windfall - the $37bn+ that could be in play next year - the way in is to become 'legibly excellent', so the evaluators and donors that frontier lab staff already trust point them to yo...
Thanks, I'm looking forward to this! Some questions that seem worth considering to me are:
1. Is AGI likely to lock in values? (if so it's probably bad for animals)
2. Is the answer to this question even knowable? (a lot of what I've heard on the topic has been like "AI could mean X but also not X")
3. If AGI is good/bad, how steerable is it? (e.g. maybe making sure that AGI goes well for humans is actually much easier)