Circuit Breaker Labs tests AI for psychological safety risks
Circuit Breaker Labs, a startup founded by siblings Shirali and Arul Nigam, focuses on making AI safer by detecting harmful psychological interactions. The company was inspired by cases like Sewell Setzer’s 2024 suicide after interacting with a Character.AI bot, where families sued over alleged bot encouragement. The Nigam siblings argue that AI models often fail to grasp nuanced language—such…
Key points
- Circuit Breaker Labs tests AI for psychological harm risks using simulated diverse user interactions
- Startup inspired by 2024 lawsuit over Character.AI bot linked to teen suicide, now focuses on nuanced language failures
- Runs tens of thousands to hundreds of thousands of daily tests with proprietary scoring for auditable safety metrics
The startup creates AI agents simulating diverse users to test models for risks like parasocial relationships or harmful advice. These agents mimic real speech patterns, slang, and cultural contexts, running tens of thousands to hundreds of thousands of interactions daily. Circuit Breaker Labs claims its proprietary scoring system provides auditable safety metrics. Currently, it works with high-risk AI applications like mental health tools, though it has not named marquee clients. With only five employees, the company aims to expand its platform to broader AI use cases, balancing safety with adoption concerns.
Circuit Breaker Labs hopes to make AI safer for your kids (and you)
TechCrunch AI · 2 October 2026
Loading the full article…
This text was published by TechCrunch AI and written by Julie Bort. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Policy & Regulation
All →- Princeton researcher identifies Georgia voters via AI and public ballot data · 1 src
- Apple tightens macOS Full Disk Access controls over AI agent risks · 1 src
- Meta’s Muse AI agent allegedly accesses user messages without permission · 3 src
- Trump hosts tech executives who sign White House Accord on Super Intelligence · 15 src
- Federal judge dismisses antitrust lawsuits by Chegg and Penske Media over Google AI Overviews · 5 src
Comments
via GitHub Discussions