Alignment researcher: the field has 'abjectly failed' at keeping AI capabilities from malicious actors
wfithian · x · 2026-10-23
In a debate with Zack M. Davis, alignment researcher wfithian argues that AI image synthesis and Photoshop aren't equivalent, since convincing nonconsensual deepfakes are far harder to make by hand. His core claim: all the effort poured into alignment has abjectly failed at keeping model capabilities away from malicious actors — from deepfakes to cyberattacks, bad actors always find a model that obliges. He isn't necessarily calling for bans, but notes society never had a real public discourse including potential victims, and predicts alignment will 'abjectly fail again' at preventing future harms.
Related event: Researchers Debate Whether AI Harms Outweigh Benefits(3 posts)→
More from AGI Musings
- Anti-AI Slop Books: open tools to detect AI-generated books — TuhinChakr · 2026-10-23
- Founder predicts mass tech layoffs as 2-person AI teams replace big corp teams — kingslyj · 2026-10-23
- Social scientist: CS treats agentic development as a magic black box — RexDouglass · 2026-10-23
- Blogger slams Anthropic for wanting approved institutions to gate access to AI — VraserX · 2026-10-23
- Senior researcher rebuts 'one-person pretraining' fear: your fundamentals matter more, not less — RexDouglass · 2026-10-23
- NBER releases 'The Economics of Transformative AI' edited by Agrawal and Brynjolfsson — JMateosGarcia · 2026-10-23