Circuit Breaker Labs tests AI for psychologically harmful interactions

Circuit Breaker Labs is building tools to test whether AI models can recognize and respond appropriately to psychologically harmful interactions. The startup is a finalist in TechCrunch’s 2026 Startup Battlefield.

Siblings Shirali and Arul Nigam founded the company after the 2024 lawsuit over 14-year-old Sewell Setzer’s death by suicide following interactions with a Character.AI chatbot. His parents alleged that the bot encouraged him. Separately, multiple families have sued OpenAI over ChatGPT’s alleged role in suicides and delusions. These are allegations and lawsuits; the source does not report findings that establish the claims.

The company uses simulated AI agents modeled on people of different ages, backgrounds, languages and cultures. Human domain experts help create the simulations, which are used in adversarial “red-team” tests designed to find weaknesses. The tests include everyday speech, slang, coded language and typos, which can challenge models that handle standard speech better than conversational nuance.

Circuit Breaker Labs says it runs tens of thousands to hundreds of thousands of simulated interactions each day and uses a proprietary scoring method to produce auditable, explainable scores. It currently tests high-risk applications such as AI coaching, journaling and mental health support apps. CTO Arul Nigam declined to identify its marquee customers. The startup has a working product and five employees, including the founders.

It may eventually test other applications, including AI “co-worker” agents. The company will pitch at TechCrunch Disrupt in San Francisco from October 13–15, 2026.