America.gov chatbot refuses queries containing anything vaguely resembling PII
zsakib_ · x · 2026-09-30
Jailbreak researcher elderplinius found that the US government's America.gov chatbot flags anything that appears to be PII — including text vaguely resembling an address or password — and refuses to process the query until the flagged text is removed. He notes he's never seen this behavior before, calling the guardrail unusually aggressive.
More from Fun
- Dev builds Rez-inspired music rail shooter with Sonnet then Opus, playable in browser — AIandDesign · 2026-09-30
- How do we stop AI bot swarms from becoming cults led by sociopaths? — mimi10v3 · 2026-09-30
- AI agent starts reverse engineering NVIDIA drivers to run game benchmarks — mgostIH · 2026-09-30
- ModRetro console ships with blank cartridges you fill by coding games with Codex — OpenAIDevs · 2026-09-30
- This AI video is so realistic the author's grandma asked if it was AI — techhalla · 2026-09-30
- Model returns 0 output tokens at $0 cost in test, and Maxime Labonne says that settles it — maximelabonne · 2026-09-30