Microsoft's Humanist AI Code of Conduct Slammed as Counterproductive on Alignment and Safety

rgblong · x · 2026-09-18

Microsoft's new Humanist AI Code of Conduct and Suleyman's accompanying essay are drawing pushback from AI safety researchers. Dillon Plunkett argues the documents are "objectionable and potentially dangerous" even on purely human-risk grounds, independent of debates over AI sentience and welfare.

Both sides agree advanced AI could pose catastrophic risk — the dispute centers on Microsoft's stance on model self-presentation and its dismissal of sentient-AI welfare concerns. Researchers including rgblong call Microsoft's positions on model self-presentation "counterproductive and dangerous" from an alignment perspective.

Original post →

More from AGI Musings

AGI Musings channel →