Some AI researchers are increasingly worried about the lack of ethical oversight, with the job often falling to peer reviewers, a stark contrast to other fields
At artificial-intelligence conferences, researchers are increasingly alarmed by what they see. Tweets: @silverjacket , @mathewjschwartz , @bhecht , @vdignum , @mkearnsupenn , and @sh_reya Tweets: Matthew Hutson / @silverjacket : “I don't want my research to succeed,” an AI grad student said, concerned about her work's impact. My latest for @NewYorker. https://www.newyorker.com/... Mathew Schwartz / @mathewjschwartz : Analysis: Who should stop unethical A.I.? At artificial-intelligence conferences, researchers are increasingly alarmed by what they see. (New Yorker) https://www.newyorker.com/... Brent Hecht / @bhecht : A good way to catch up on developments in the academic community around tech ethics. I like the focus on HCI's leadership in this space (cc @sigchi). The article also does a good job describing the most jarring conversation I've ever had in a conference poster session! https://twitter.com/... Virginia Dignum / @vdignum : Good read even if misleading title: There is no unethical #AI (in the same way that there is no ethical AI). There are unethical approaches and uses of AI. #responsibleAI #AIethics Who Should Stop Unethical A.I.? https://www.newyorker.com/... via @NewYorker Michael Kearns / @mkearnsupenn : Extensive and balanced article on ethical challenges in AI/ML, nicely done @SilverJacket https://twitter.com/... Shreya Shankar / @sh_reya : Good read on AI research <-> ethics tensions. It's terrifying to think about what the equivalent of the Tuskegee Study in AI research will be. Or, where is the “Nuremberg Code” for ethically questionable forecasting tasks? https://twitter.com/...
Context & Ripple Effects
The New Yorker piece lands two weeks after Timnit Gebru's dismissal from Google, which Wired framed as proof of tech's ability to co-opt and minimize internal criticism of its AI systems — so the question of who checks unethical AI had just lost its most visible inside answer.
The article's framing borrows from older research-governance precedents: the [[entity/the-“nuremberg-code|Nuremberg Code]] and the Tuskegee Study, which produced formal review boards in medicine and other fields. Earlier coverage had already argued that ethics boards can't implement good tech on their own, only raise awareness and force self-criticism — this piece extends that argument to a field where no such boards exist at all.
First-order effects
- Peer reviewers at AI conferences are now the de facto ethics infrastructure: researchers like Shreya Shankar and Brent Hecht are asking whether volunteer reviewers, with no training or mandate for it, should carry a burden other fields assign to dedicated institutions.
- Grad students doing the work feel the weight directly — one quoted in the piece says she doesn't want her own research to succeed over concerns about its impact.
Second-order effects
- With internal critics like Gebru shown to be removable, the checking function migrates outward to conferences and journals, raising the stakes of whether venues adopt explicit ethics-review criteria or leave it ad hoc.
- Companies whose systems get flagged by reviewers face a legitimacy problem they can't resolve by hiring ethicists alone — the earlier Slate argument suggests boards without implementation power become moral cover rather than oversight.
Third-order effects
- If the pattern holds, AI research may be pushed toward formalizing its own review institutions modeled on the medical-ethics lineage the article invokes — or toward external regulation, which related coverage notes is lagging even as security researchers warn about large-scale destruction potential.
- The alternative is a field where ethical judgment stays distributed across unpaid volunteers while commercial incentives accelerate — a structure that echoes the burnout dynamics described in an ML scientist's account of an industry shaken by ChatGPT.
The trend: AI is confronting the same governance inflection medicine faced after Nuremberg and Tuskegee — whether a fast-commercializing research field builds formal ethical review before regulators impose it.