Tegmark's 6 Moral Questions Challenge AI Researchers to Define Red Lines

Max Tegmark, an MIT AI professor, and Meia Chita-Tegmark, a psychology researcher, published a piece on LessWrong urging AI researchers to actively practice moral reasoning. They argue that most ethical failures come not from lacking principles but from failing to activate them. The article presents six guiding questions: Do you have red lines? Have you written them down and shared them? Have you resisted moral disengagement? (the remaining three are implied but build on these). They cite Daniel Kokotajlo, who walked away from nearly $2M in OpenAI equity to uphold his red lines, and the ongoing clash between Anthropic and the Pentagon as real-world examples. The authors also draw on Albert Bandura's moral disengagement mechanisms—like displacement and diffusion of responsibility—to explain how researchers can rationalize harmful actions. They reference Hannah Arendt's concept of 'just doing my job' and Tom Lehrer's satirical song about Wernher von Braun to underscore how accountability can fade in large organizations.

The piece then warns that AI researchers are especially vulnerable because their work is distributed across teams, making collective outcomes feel less personal. The authors provide a link to list personal red lines (shared only with permission) and encourage social accountability to prevent the 'boiling frog' effect of incremental ethical compromise. Although the full six questions are not all listed in the provided excerpt, the core message is clear: researchers must proactively define, document, and defend their moral boundaries against external pressures from companies, peers, and market forces. The article serves as both a call to action and a practical toolkit for those building what they call 'the most powerful technology ever.' It emphasizes that individual choices matter and that moral muscles must be exercised regularly to avoid drift.