December 30, 2024

I may have been effective altruism’s (EA) first major whistleblower. Given it was in relation to a major, seminal AI safety initiative and EA was a major steward of AI safety, I may have been one of the first whistleblowers in the AI safety space, as well. See my related company’s statement on the matter.

If the incident had not occurred, likely hundreds of thousands of hours and millions of dollars would have otherwise been put to much better use improving the world. 

I reported the incident to the leadership of a central EA organization. I have not discussed the incident publicly in detail due to the negative impact on me and others, an NDA, my uncertainty over the actors involved, and fear of whistleblower retaliation. 

Fortunately, most of the bad actors involved were eventually removed from the broad movement (insofar as that is possible). And I’ve learned a lot more about what actually happened over the years. 

Unfortunately, EA/AI safety leadership did not appear to learn much from the incident, as the Machiavellian pattern I outlined to them was repeated years later in Sam Bankman-Fried (SBF). The FTX fraud was foreseeable. Indeed, I spent tens of thousands of hours trying to prevent such a case from arising, only to have my efforts be mostly ignored by EA/AI safety leadership.*

I think there’s a 30% chance that if EA/AI safety leadership had shown more wisdom with the proto-SBF individual that I blew the whistle on, then the FTX fraud would not have later happened. It is also possible that a marginally more honest culture could have been embedded in some of the frontier AI companies.  

Over the years, beyond my incident I’ve also tracked dozens of other scandals in the broad EA and AI safety communities. I had hoped that key leaders would have begun to understand their role as stewards of these communities somewhat better. But that doesn’t seem to have happened. 

For those who are curious, here are my proposed EA Network Governance System and EA Network Funding Guidelines.

 

* For 10+ years I offered personal advisory boards and fractional executive support to help influential leaders–especially EAs–think and execute more efficiently, effectively, and ethically. We explicitly designed the Upgrade Program (UP) to attract, support, and guide the altruistic high-achieving type–Sam Bankman-Fried, Will MacAskill, Elon Musk, Sam Altman, Helen Toner, and others–knowing how their stress, isolation, overwhelm, and need for privacy would occasionally lead to poor decision-making that would massively affect us all. I also did a fair amount of other behind-the-scenes work to improve the integrity of the EA ecosystem in other ways. None of this was funded or supported by EA leadership.