Cullen_OKeefe

• Cullen_OKeefe 15 Aug 2022 17:53 UTC

  2 points

  in reply to: John_Maxwell’s comment on: Rhetor­i­cal Abus­abil­ity is a Poor Counterargument

  (Just flagging that this is very related to the discussion in the first part of Reasons and Persons, and for the reasons presented therein I don’t think it’s a decisive argument against consequentialism as criterion of rightness.)

• Cullen_OKeefe 8 Aug 2022 20:32 UTC

  7 points

  in reply to: Kaleem’s comment on: What should we call the Har­vard EA office ?!?!?!?

  (I agree)

• Cullen_OKeefe 8 Aug 2022 20:21 UTC

  2 points

  in reply to: Sharmake’s comment on: What should we call the Har­vard EA office ?!?!?!?

  I would note that advocating for improving utility (a core EA concept!) is not the same thing as utilitarianism.

• Cullen_OKeefe 8 Aug 2022 20:09 UTC

  4 points

  in reply to: Kaleem’s comment on: What should we call the Har­vard EA office ?!?!?!?

  Plausibly not!

• Cullen_OKeefe 8 Aug 2022 19:56 UTC

  2 points

  in reply to: RyanCarey’s comment on: What should we call the Har­vard EA office ?!?!?!?

  For a journal! But surely an office can have the same name as a journal? It’s just a Latin word!

• Cullen_OKeefe 8 Aug 2022 19:30 UTC

  2 points

  on: What should we call the Har­vard EA office ?!?!?!?

  “Utilitas.” Consequentialist vibes + similarity to Harvard’s motto (sorry MIT!).

How tech­ni­cal safety stan­dards could pro­mote TAI safety

Jan Leike: On the wind­fall clause

• Cullen_OKeefe 4 Aug 2022 20:49 UTC

  4 points

  in reply to: Lorenzo’s comment on: Open Thread: June — Septem­ber 2022

  Ah! I was ignorant of this. Thank you!

• Cullen_OKeefe 4 Aug 2022 19:17 UTC

  3 points

  on: Open Thread: June — Septem­ber 2022

  @Lizka, a cool forum feature would be a Google Doc/​.docx → Forum Post converter. There are free .docx → Markdown converters online, but they’re inconvenient for footnotes and some of the other EAF features.

Law-Fol­low­ing AI 4: Don’t Rely on Vi­car­i­ous Liability

• Cullen_OKeefe 29 Jul 2022 0:00 UTC

  7 points

  on: Effec­tive Altru­ism for Muslims

  Congrats! Super exciting.

• Cullen_OKeefe 24 Jul 2022 2:50 UTC

  3 points

  in reply to: Luke Chambers’s comment on: Law-Fol­low­ing AI 3: Lawless AI Agents Un­der­mine Sta­bi­liz­ing Agreements

  Thanks a ton for your substantive engagement, Luke! I’m sorry it took so long to respond, but I highly value it.

  Law doesn’t always reflect morals we want it to. In the 1940s, an AI system under this type of governance would have turned any Jews it found in to the Nazis because it was the lawful thing to do, despite the fact it is objectively to a human the wrong thing to do. Further examples are turning in escaped slaves it encountered, fully collaborating with Gestapo hunting partisans, full collaboration with Russians seizing Ukrainian territory, and more. These are all extreme examples sprinkled throughout the past—but the point is we don’t know what the future holds. Currently here in the UK the government are throwing lots of effort and resources into passing laws to restrict basic protest rights under the Policing and Crime Bill. It’s not impossible in 30 years time for the UK (or anywhere else) to be a police state or authoritarian regime. An AI rigidly sticking to the law would be misaligned morally. You did focus on this when discussing a balance of outcomes so it’s not so much a weakness as an area we need to explore much more.

  Yeah, definitely agree that this is tricky and should be analyzed more (especially drawing on the substantial existing literature about moral permissibility of lawbreaking, which I haven’t had the time to fully engage in).

  For Common Law systems such as the UK (and I believe USA? Please correct if wrong), the AI utilising case law for its moral alignment would change faster than we could reasonably foresee. Just look at the impact R v Cartwright (1569) had on Shanley v Harvey (1763) by not mentioning the race of the slave who was to be considered a person and freed because he was on English soil. For Civil Law systems such as those in mainland Europe this would be a much easier idea, but Common Law systems would need a bit more finesse. Or just not including case law and relying on legislation, which could be a bit more barebones.

  Yeah, I do think there’s an interesting thing here where LFAI would make apparent the existing need to adopt some jurisprudential stance about how to think about the evolution of law, and particularly of predicted changes in the law. As an example of how this already comes up in the US, judges sometimes regard higher courts’ precedents as bad law, notwithstanding the fact that the higher court has not yet overruled it. The addition of AI into the mix—as both a predictor of and possible participant in the legal system, as well as a general accelerator of the rate of societal change—certainly threatens to stretch our existing ways of thinking about this. This is also why I’m worried about asymmetrical use of advanced AI in legal proceedings. See footnote 6.

  (And yes, the US^[1] is also common law. :-) )

  Different countries have different laws. Some companies in the USA don’t (or can’t) operate in the EU and in the UK because GDPR affords Europeans more data rights than the USA does on its own citizens, which means that for data processing companies which even allow access to their websites from European jurisdiction would face large fines for doing to EU resident data what they regularly do to US data with no problem. I can see this being an issue with moral alignment. When we say for AI to follow law—whose law? US rights and freedoms laws imposed on European AI would be a collossal backstep for human and civil rights in Europe not seen since WW2, whereas European copyright and patent laws imposed to US AI would be a huge backstep for the US (don’t get me started on the EU’s patent laws!). If the AI just follows the laws where the AI is based, it will be difficult to operate internationally for all except physical robot systems which could follow the same rules as people currently do without much effort given to changing its thinking as soon as it passed physical boundaries or borders. Perhaps we could create an international agreement on core concepts such as the Universal Declaration of Human Rights (UDHR), the International Covenant on Civil and Political Rights (ICCPR), the International Covenant on Economic, Social, and Cultural Rights (ICCPR), and the European Convention on Human Rights (ECHR). We did it for unifying laws about rights to avoid another holocaust—why not about alignment to avoid a similar, yet potentially larger, catastrophe?

  Definitely agree. I think the practical baby step is to develop the capability of AI to interpret and apply any given legal system. But insofar as we actually want AIs to be law-following, we obviously need to solve the jurisdictional and choice of law questions, as a policy matter. I don’t think we’re close to doing that—even many of the jurisdictional issues in cyber are currently contentious. And as I think you allude to, there’s also a risk of regulatory arbitrage, which seems bad.

  ---

  1. ↩︎

     Except the civil law of Louisiana, interestingly.

• Cullen_OKeefe 24 Jul 2022 2:35 UTC

  2 points

  in reply to: aogara’s comment on: Law-Fol­low­ing AI 1: Se­quence In­tro­duc­tion and Structure

  Thanks Aidan! I do regret this not being as discoverable as optimal, but I’m kinda publishing this as I’m writing it. Maybe I’ll do a white paper as I refine the central points.

  We should chat about this sometime. I also didn’t see your discussion of law; seems like we have untapped shared interests!

Pile of Law and Law-Fol­low­ing AI

• Cullen_OKeefe 10 Jul 2022 0:57 UTC

  2 points

  on: Should I in­vest my run­way?

  You might find this article helpful: https://​​80000hours.org/​​2015/​​10/​​common-investing-mistakes-in-the-effective-altruism-community/​​

• Cullen_OKeefe 18 Jun 2022 1:54 UTC

  4 points

  in reply to: Charles He’s comment on: Jobs at EA-or­ga­ni­za­tions are over­paid, here is why

  (For context, I’m a lawyer who’s worked a lot on issues related to antitrust and AI safety.)

  These entities are funded by very close and coordinated sources of funds

  There is such a thing as the “single entity defense,” but it almost certainly would not apply in this context. The orgs are legally and functionally separate entities.

  The “cartel” actually has to create economic harm, and this might require a natural marketplace for employees that is being undermined. If the funders and all the execs created the entire marketplace (e.g. think of many EA cause areas), that’s harder to say they are unfairly benefitting. . . . The EA community or funders gathered together and demanded lower/​higher salaries or other norms.

  The best theory would be that employees are being harmed by anticompetitive fixing of the prices of their employment. That the funders or “community” want the salaries to be lower does not negate the impact on the workers, and so is not relevant to the claim.

  However, it may be less legally risky (in the US) for funders to impose (common) salary caps on the downstream orgs that they fund as a condition of funding. Major cases on-point: State Oil Co. v. Khan, 522 U.S. 3 (1997); Continental T.V., Inc. v. GTE Sylvania, Inc., 433 U.S. 36 (1977); Leegin Creative Leather Products, Inc. v. PSKS, Inc., 551 U.S. 877 (2007).

• Cullen_OKeefe 18 Jun 2022 1:45 UTC

  2 points

  on: An­titrust law

  NB: The formatting of the Brundage article is intentional as the first five authors had equal contributions.

• Cullen_OKeefe 17 Jun 2022 21:53 UTC

  12 points

  in reply to: Aleks_K’s comment on: Jobs at EA-or­ga­ni­za­tions are over­paid, here is why

  An agreement of the type contemplated in the OP (as I understand it) would very likely be treated as anticompetitive “per se,” and therefore not require any empirical inquiry into the effects on the “market” (however defined).

• Cullen_OKeefe 17 Jun 2022 20:56 UTC

  6 points

  in reply to: Hauke Hillebrandt’s comment on: Cul­len_OKeefe’s Shortform

  by the same logic, wouldn’t you also want to overexpose yourself to e.g. publicly traded real estate because most of it isn’t public?

  Yes, I think that is true as well!

  If true, and if most passive (institutional) investors aren’t sufficiently exposed to PE (or real estate), wouldn’t that suggest that the market undervalues this asset class and you can beat the market by investing in it?

  This is what I was getting at by noting that PE should also be higher beta, so the risk-adjusted returns should not be higher!

  Honest question, I haven’t thought this through very well, but something still feels counterintuitive that you could create a better passive global market portfolio...

  Hm, very possible I am thinking about this wrong, but it’s pretty intuitive to me. After all, isn’t the “true” global market all investable assets, not just publicly traded assets? It’s the same reason why the Nasdaq Composite is a poor stand-in for the global stock market: for a variety of reasons, it only has a certain subset of publicly traded equities, and other baskets of equities have different risk-return profiles.