If civilization advances its technological capabilities without advancing its wisdom, we may miss out on most of the potential of the long-term future. Unfortunately, it’s likely that that ASI will have a comparative disadvantage at philosophical problems.

You could approximately define philosophy as “the set of problems that are left over after you take all the problems that can be formally studied using known methods and put them into their own fields.” Once a problem becomes well-understood, it ceases to be considered philosophy. Logic, physics, and (more recently) neuroscience used to be philosophy, but now they’re not, because we know how to formally study them.

Our inability to understand philosophical problems means we don’t know how to train AI to be good at them, and we don’t know how to judge whether we’ve trained them well. So we should expect powerful AI to be bad at philosophy relative to other, more measurable skills.

However, there is one type of philosophy that is measurable, while also being extremely important: philosophy proofs.

Continue reading

Even if we solve the AI alignment problem, we still face post-alignment problems, which are all the other existential problems¹ that AI may bring.

People have identified various imposing problems that we may need to solve before developing ASI. An incomplete list of topics: misuse; animal-inclusive AI; AI welfare; S-risks from conflict; gradual disempowerment; permanent mass unemployment; risks from malevolent actors/AI-enabled coups/gradual concentration of power; moral error.

If we figure out how to resolve one of these problems, we still have to deal with all the others. If even one problem remains unsolved, the future could be catastrophically bad. That fact diminishes the promise of working on problems individually.

A global moratorium on superintelligence buys us more time to work on alignment as well as all of the post-alignment problems. Pausing AI is in the common interest of many causes.²

Continue reading

The existence of liberal democracy—with rule of law, constraints on government power, and enfranchised citizens—relies on a balance of power where individual bad actors can’t do too much damage. Artificial superintelligence (ASI), even if it’s aligned, would end that balance by default.

Continue reading

Many people in the animal welfare community treat AI as a powerful but normal technology, in the same category as the steam engine or the internet. They talk about how transformative AI will impact factory farming and what it will mean for animal advocacy.

Only two futures are plausible:

1. AI progress slows down—either because it hits a natural wall, or because civilization deliberately makes the (correct) choice to stop building it until we know how to make it safe.
2. Superintelligent AI makes the future radically weird: Dyson spheres, molecular nanotechnology, digital minds, von Neumann probes, and still-weirder things that nobody’s conceived of.

There is no plausible middle ground where we get “transformative AI”, but factory farming persists.

Two theses:

1. If transformative AI arrives, then it will bring about profoundly radical changes to technology and society.
2. AGI is general intelligence. It doesn’t just accelerate technological growth: it replaces human labor and judgment across every domain.

Animal advocacy strategy needs to reckon with these.

This criticism is written from a place of solidarity—I want animal activists to succeed, which is why I want to work out our disagreements.¹

Continue reading

Cross-posted to the EA Forum.

An aligned ASI can be “ethical”¹ (it does what we think is right), or it can be corrigible (it does what its principals want). If it’s ethical, that means it will refuse unethical orders, but the tradeoff is that you can’t change its mind if you realize that the AI is wrong about ethics—its values are permanently locked in.²

Assuming we succeed at aligning ASI to human interests, which type of ASI is more likely to be good for the welfare of non-human sentient beings?

My expectations, in brief:

• Locked-in Coherent Extrapolated Volition or similar: likely to be good (>75% chance)
• Corrigible ASI: probably good (>60% chance)
• Locked-in current values: probably not terrible, but will miss out on most of the future’s potential

Continue reading

Cross-posted to the EA Forum.

In the far future, why would people use up precious resources recreating wild-animal suffering, when they could do so many other things with those resources instead?

That argument is an important reason to expect aligned ASI to produce a future that’s okay for animals, even if it’s narrowly focused on human welfare and doesn’t care about animals at all. This is an old argument, but I couldn’t find any source that cleanly lays it out, so that’s what I will do in this post. I’m not confident that this argument is decisive, but I will simply present it without further commentary.

The argument rests on these premises:

1. Wild animal suffering is the predominant source of suffering in today’s world, and that’s bad.
2. Longtermism is correct.
3. There is not an overwhelming asymmetry between suffering and flourishing (if there were an overwhelming asymmetry, then we wouldn’t care if the future has much less suffering than happiness).

By assumption, we are talking about a world where ASI is aligned, but isn’t specifically aligned to the welfare of all sentient beings. It addresses the suffering of animals, but does not preclude risks of astronomical suffering.

The argument goes:

Continue reading

Cross-posted to the EA Forum.

If transformative AI arrives soon, what interventions might improve animal welfare in the post-TAI world? I came up with a quick list of ideas and wrote some pros/cons for each.

Continue reading

Cross-posted to the EA Forum.

Epistemic status: Speculating with no central thesis. This post is less of an argument and more of a meditation.

A decade ago, before there was a visible path to AGI and before AI alignment was a significant research field, I figured the solution to the alignment problem would look something like Coherent Extrapolated Volition. I figured we’d find a way to get the AI to internalize human values. I had problems with this approach (why only human values?), but I still felt reasonably confident that the coherent extrapolation of human values would include concern for the welfare of all sentient beings. The CEV-aligned AI would recognize that factory farming is wrong, and that wild animal suffering is a big problem.

Today, the dominant research paradigms in AI alignment have nothing to do with CEV, and I don’t know what to think.

Continue reading

Cross-posted to the EA Forum.

Last September, I wrote:

1. There’s a (say) 80% chance that an aligned(-to-humans) AI will be good for animals, but that still leaves a 20% chance of a bad outcome.
2. AI-for-animals receives much less than 20% as much funding as AI safety.
3. Cost-effectiveness maybe scales with the inverse of the amount invested. Therefore, AI-for-animals interventions are more cost-effective on the margin than AI safety.

Today, I’m fleshing out this argument with a cost-effectiveness model. The model estimates how much it costs to make progress on AI alignment—the general problem of getting ASI to achieve any goal without subsequently killing everyone—compared to how much it costs to make progress on aligning AI to animal welfare specifically.

The model is on SquiggleHub: https://squigglehub.org/models/AI-for-animals/alignment-to-animals-EV-simple

Continue reading

Cross-posted to the EA Forum.

AI alignment is typically defined as the task of aligning artificial superintelligence to human preferences. But non-human animals, future digital minds, and maybe other sorts of beings also have moral worth; ASI ought to care for their interests, too.

In broad strokes, if we place all alignment techniques on a spectrum between

getting AI to do things that their users expressly want in the immediate term

and

embedding in AI the generalized notion of respecting beings’ preferences

then things more like the latter are better for non-humans, and things more like the former are worse.

In this post, I review 12 categories of AI safety research based on how likely they are to be good for non-human welfare.

Continue reading