## LESSWRONGLW

Ben Cottier

Working on modelling beliefs about AI risk in spare time

Sorted by New

# Wiki Contributions

What are the biggest current impacts of AI?

Personal AI assistants seem to have one of the largest impacts (or at least "presence") mainly due to the number of users. The impact per person seems small - making life slightly more convenient and productive, maybe. Not sure if there is actually much impact on productivity. I wonder if there is any research on this. I haven't looked into it at all.

Relatedly, chatbots are certainly used a lot, but I'm uncertain about its current impacts beyond personal entertainment and wellbeing (and uncertain about the direction of the impact on wellbeing).

What 2026 looks like has a few relevant facts on the current impacts, and interesting speculation about the future impacts of personal assistants and chatbots. E.g. facts:

"in China in 2021 the market for chatbots is $420M/year, and there are 10M active users. This article claims the global market is around$2B/year in 2021 and is projected to grow around 30%/year."

I don't feel surprised by those stats, but I also hadn't really considered how big the market is.

Modeling the impact of safety agendas

Nice! A couple things that this comment pointed out for me:

1. Real time is not always (and perhaps often not) the most useful way to talk about timelines. It can be more useful to talk about different paths, or economic growth, if that's more relevant to how tractable the research is.
2. An agenda doesn't necessarily have to argue that its assumptions are more likely, because we may have enough resources to get worthwhile expected returns on multiple approaches.

Something that's unclear here: are you excited about this approach because you think brain-like AGI will be easier to align? Or is it more about the relative probabilities / neglectedness / your fit?

AI learns betrayal and how to avoid it

I'm excited about this project. I've been thinking along similar lines about inducing a model to learn deception, in the context of inner alignment. It seems really valuable to have concrete (but benign) examples of a problem to poke at and test potential solutions on. So far there seem to be less concrete examples of deception, betrayal and the like to work with in ML compared to say, distributional shift, or negative side effects.

AI learns betrayal and how to avoid it

Previous high level projects have tried to define concepts like "trustworthiness" (or the closely related "truthful") and motivated the AI to follow them. Here we will try the opposite: define "betrayal", and motivate the AIs to avoid it.

Why do you think the betrayal approach is more tractable or useful? It's not clear from the post.

Clarifying some key hypotheses in AI alignment

To your first point - I agree both with why we limited the scope (but also, it was partly just personal interests), and that there should be more of this kind of work on other classes of risk. However, my impression is the literature and "public" engagement (e.g. EA forum, LessWrong) on catastrophic AI misuse/structural risk is too small to even get traction on work like this. We might first need more work to lay out the best arguments. Having said that, I'm aware of a fair amount of writing which I haven't got around to reading. So I am probably misjudging the state of the field.

To your second point - that seems like a real crux and I agree it would be good to expand in that direction. I know some people working on expanded and more in-depth models like this post. It would be great to get your thoughts when they're ready.

Clarifying some key hypotheses in AI alignment

It's great to hear your thoughts on the post!

I'd also like to see more posts that do this sort of "mapping". I think that mapping AI risk arguments is too neglected - more discussion and examples in this post by Gyrodiot. I'm continuing to work collaboratively in this area in my spare time, and I'm excited that more people are getting involved.

We weren't trying to fully account for AGI timelines - our choice of scope was based on a mix of personal interest and importance. I know people currently working on posts similar to this that will go in-depth on timelines, discontinuity, paths to AGI, the nature of intelligence, etc. which I'm excited about!

I agree with all your points. You're right that this post's scope does not include broader alternatives for reducing AI risk. It was not even designed to guide what people should work on, though it can serve that purpose. We were really just trying to clearly map out some of the discourse, as a starting point and example for future work.

Conditions for Mesa-Optimization

A system capable of reasoning about optimization is likely also capable of reusing that same machinery to do optimization itself

"A system capable of reasoning about photosynthesis is likely also capable of reusing that same machinery to do photosynthesis itself." Nope.

"A system capable of reasoning about arithmetic is likely also capable of reusing that same machinery to do arithmetic itself". Maybe? The rules of arithmetic can be reused, but the machinery to reason abstractly about arithmetic is probably different to the machinery to run a specific calculation, especially with a learned model with lots of free parameters, like a neural network.

Maybe optimisation is not like the above examples because it is so generic? Or I misunderstood the claim.

Do mesa-optimizer risk arguments rely on the train-test paradigm?

Thanks. I think I understand, but I'm still confused about the effect on the risk of catastrophe (i.e. not just being pseudo-aligned, but having a catastrophic real-world effect). It may help to clarify that I was mainly thinking of deceptive alignment, not other types of pseudo-alignment. And I'll admit now that I phrased the question stronger than I actually believe, to elicit more response :)

I agree that the probability of pseudo-alignment will be the same, and that an unrecoverable action could occur despite the threat of modification. I'm interested in whether online learning generally makes it less likely for a deceptively aligned model to defect. I think so because (I expect, in most cases) this adds a threat of modification that is faster-acting and easier for a mesa-optimizer to recognise than otherwise (e.g. human shutting it down).

If I'm not just misunderstanding and there is a crux here, maybe it relates to how promising worst-case guarantees are. Worst-case guarantees are great to have, and save us from worrying about precisely how likely a catastrophic action is. Maybe I am more pessimistic than you about obtaining worst-case guarantees. I think we should do more to model the risks probabilistically.

Deceptive Alignment

In the limit of training on a diverse set of tasks, we expect joint optimization of both the base and mesa- objectives to be unstable. Assuming that the mesa-optimizer converges towards behavior that is optimal from the perspective of the base optimizer, the mesa-optimizer must somehow learn the base objective.

Joint optimization may be unstable, but if the model is not trained to convergence, might it still be jointly optimizing at the end of training? This occurred to me after reading https://arxiv.org/abs/2001.08361 which finds that "Larger models are significantly more sample-efficient, such that optimally compute-efficient training involves training very large models on a relatively modest amount of data and stopping significantly before convergence." If convergence is becoming less common in practical systems, it's important to think about the implications of that for mesa-optimization.