20% of the UK Milburn Review was AI-written, sparking disclosure norms debate
birchlse · x · 2026-11-19
Daisy Christodoulou flagged that 20% of the Milburn Review was written with AI, asking what the norms now are. She suggests heavy AI use may warrant disclosure of prompts (e.g. "turn these bullet points about kids into emotive paragraphs"), admits her instinctive dislike may be snobbery, and has an article coming in Google's AI Policy Perspectives Substack.
Related event: UK Milburn Review found to be 20% AI-generated(2 posts)→
More from Safety
- rasbt mocks EU AI labels: even Pixar-style animated ads get tagged — rasbt · 2026-11-19
- Nexus founder Thom Wolf explains why he signed the letter accusing Anthropic of capturing the AI safety ecosystem — stanfordnlp · 2026-11-19
- Xint to disclose five critical nginx CVEs (CVSS 9.2) found via AI vulnerability discovery — dyn___ · 2026-11-19
- Same models, more bugs: Xint's harness finds 5 critical nginx bugs, 3 never found before — dyn___ · 2026-11-19
- How Much Inbox Access Should an Agent Get? Debating Fine-Grained Scopes — yenkel · 2026-11-19
- Erik Hoel: engineering non-conscious AI is a civilization-level requirement — erikphoel · 2026-11-19