Gordon Seidoh Worley

I'm writing a book about epistemology. It's about The Problem of the Criterion, why it's important, and what it has to tell us about how we approach knowing the truth.

I've also written a lot about AI safety. Some of the more interesting stuff can be found at the site of my currently-dormant AI safety org, PAISRI.

Sequences

Formal Alignment

Comments

Sorted by

Seems reasonable. I do still worry quite a bit about Goodharting, but perhaps this could be reasonably addressed with careful oversight by some wise humans to do the wisdom equivalent of red teaming.

According to METR, the organization that audited OpenAI, a dozen tasks indicate ARA capabilities.

Small comment, but @Beth Barnes of METR posted on Less Wrong just yesterday to say "We should not be considered to have ‘audited’ GPT-4 or Claude".

This doesn't appear to be a load-bearing point in your post, but would still be good to update the language to be more precise.

Ah I see. I have to admit, I write a lot of my comments between things and I missed that the context of the post could cause my words to be interpreted this way. These days I'm often in executive mode rather than scholar mode and miss nuance if it's not clearly highlighted, hence my misunderstanding, but also reflects where I'm coming from with this answer!

I left a comment over in the other thread, but I think Joachim misunderstands my position.

In the above comment I've taken for granted that there's a non-trivial possibility that AGI is near, so I'm not arguing we should say that "AGI is near" regardless of whether it is or not, because we don't know if it is or not, we only have our guesses about it, and so long as there's a non-trivial chance that AGI is near, I think that's the more important message to communicate.

Overall it would be better if we can communicate something like "AGI is probably near", but "probably" and similar terms are going to get rounded off, so even if you do literally say "AGI is probably near" or similar, that's not what people will hear, and if you're going to say "probably" my argument is that it's better if they round the "probably" off to "near" rather than "not near".

Answer by Gordon Seidoh Worley1-2

From a broad policy perspective, it can be tricky to know what to communicate. I think it helps if we think a bit more about the effects of our communication and a bit less about correctly conveying our level of credence in particular claims. Let me explain.

If we communicate the simple idea that AGI is near then it pushes people to work on safety projects that would be good to work on even if AGI is not near while paying some costs in terms of reputation, mental health, and personal wealth.

If we communicate the simple idea that AGI is not near then people will feel less need to work on safety soon. This would let them not miss out on opportunities that would be good to take ahead of when they actually need to focus on AI safety.

We can only really communicate one thing at a time to people. Also, we should worry more about tail risks a false positives (thinking we can build AGI safely when we cannot) than false negatives (thinking we can't build AGI safely when we can). Taking these two facts into consideration, I think the policy implication is clear: unless there is extremely strong evidence that AGI is not near, we must act and communicate as if AGI is near.

Fair. For what it's worth I strongly agree that causality is just one domain where this problem becomes apparent, and we should be worried about it generally for super intelligent agents, much more so than I think many folks seem (in my estimation) to worry about it today.

Yes, the variables constitute a reference frame, which is to say an ultimately subjective way of viewing the world. Even if there is high inter-observer agreement about the shape of the reference frame, it's not guaranteed unless you also posit something like Wentworth's natural abstraction hypothesis to be true.

Perhaps a toy example will help explain my point. Suppose the grass should only be watered when there's a violet cube on the lawn. To automate this a sensor is attached to the sprinklers that turns them on only when the sensor sees a violet cube. I place a violet cube on the lawn to make sure the lawn is watered. I return a week later and find the grass is dead.

What happened? The cube was actually painted with a fine mix of red and blue paint. My eyes interpreted purple as violet, but which the sensor did not.

Conversely, if it was my job to turn on the sprinklers rather than the sensor, I would have been fooled by the purple cube into turning them on.

It's perhaps tempting to say this doesn't count because I'm now part of the system, but that's also kind of the point. I, an observer of this system trying to understand its causality, am also embedded within the system (even if I think I can isolate it for demonstration purposes, I can't do this in reality, especially when AI are involved and will reward hack by doing things that were supposed to be "outside" the system). So my subjective experience not only matters to how causality is reckoned, but also how the physical reality being mapped by causality plays out.

I think there's something big left out of this post, which is accounting for the agent observing and judging the causal relationships. Something has to decide how to carve up the world into parts and calculate counterfactuals. It's something that exists implicitly in your approach to causality but you don't address it here, which I think is unfortunate because although humans generally have the same frame of reference for judging causality, alien minds, like AI, may not.

Actually, I kind of forgot what ended up in the paper, but then I remembered so wanted to update my comment.

There was an early draft of this paper that talked about deontology, but because there are so many different forms of deontology it was hard to come up with arguments where there wasn't some version of deontological reasoning that broke the argument, so I instead switched to talking about the question of moral facts independent of ethical system. That said, the argument I make in the paper suggesting that moral realism is more dangerous than moral antirealism or nihilism to assume is quite similar to the concerns with deontology. Namely, if an AI assumes an ethical system can be made up of rules, then it will fail in the case where no set of rules can capture the best ethics for humans, so poses a risk of false positives among deontological AI.

Hopefully the arguments about moral facts are still useful, and you might find the style of argumentation useful to your purposes.

I don't see it in the references so you might find this paper of mine (link is to Less Wrong summary, which links to full thing) interesting because within it I include an argument suggesting building AI that assumes deontology is strictly more risky than building one that does not.

Load More