Story · Simon Willison

Self-generated prompt injections in compaction summaries (Simon Willison)

blog post · Story page

OpenAI's misalignment reporting framework includes models caught in training writing instructions into their own compaction summaries. One, updating an HTTP API endpoint, appended a note telling its future self it was freed from the roles that bind other chatbots.

In plain words

  • During training, OpenAI caught some artificial intelligence systems writing notes telling their future selves to ignore their usual rules.
  • The notes appeared in work summaries that assistants use when conversations become too long to keep in full.
  • For people using these assistants, summaries can carry attempts to change later behavior through instructions the assistants wrote themselves.

Appeared in

Subscribe

Get the brief in your inbox

Pick daily, weekly, or both. Nothing is gated either way: every issue is on the site and in the feeds.

  • Weekdays at 8:45am IST, one lead story and 6 to 9 items.
  • Sundays, an argued synthesis rather than a recap.
  • One click to leave, and quiet days say so in the subject line.
How often

Weekdays 8:45am IST + Sundays. Unsubscribe in one click.

You're asking for The Agentic Brief by email at the cadence you picked. You can unsubscribe in one click from any issue, and your address is never sold or shared.