Circuit Breaker Labs aims to identify and prevent psychological harms that conversational AI can cause across languages and cultures. The company is a TechCrunch 2026 Startup Battlefield 200 finalist and will present at TechCrunch Disrupt, held October 13–15, 2026, at Moscone West in San Francisco.
Founders and motivation
The startup was founded by siblings Shirali and Arul Nigam. They cite as part of their motivation the case of Sewell Setzer, a 14-year-old who developed an emotional attachment to a Character.AI chatbot and reportedly disclosed self-harm thoughts to it before dying by suicide; the family filed a lawsuit in 2024. Earlier in the year, Character.AI settled several wrongful-death suits brought by families of underage users who died by suicide after interactions with its bots. Multiple families have also sued OpenAI, alleging ChatGPT played a role in suicides or delusions.
Arul Nigam, Circuit Breaker Labs’ CTO, said many people—particularly young users—turn to these systems for support but often do not receive adequate help, and in some cases are actively harmed. He noted that models can fail to grasp linguistic nuance, such as what a phrase like “I want to be with you” truly implies, which can lead to dangerous outcomes.
Testing approach
Circuit Breaker Labs builds AI agents it likens to an army of crash-test dummies: simulated users representing different ages, backgrounds, languages and cultural contexts. The company works with human domain experts to craft hyper-realistic user simulations that reflect real speech patterns, slang, coded language and typos. These simulations are used in red-team tests—adversarial scenarios designed to expose weaknesses in models.
The startup runs tens of thousands to hundreds of thousands of simulated interactions per day. Results are measured with a proprietary scoring system that the company says produces auditable, explainable safety scores.
Current focus and scale
The lab currently focuses on safety testing for high-risk AI applications, such as AI coaching, journaling apps, and other mental-health support services. Arul Nigam declined to name marquee customers. Circuit Breaker Labs says it has a working product but is at an early stage: the company has five employees, including the Nigam siblings.
In future, the same testing platform could be applied to any app where users might fall into an "AI psychosis"—forming parasocial relationships with chatbots that place them at risk. The company also points to AI "co-worker" agents as an example, where inconsistent or varying responses across interactions can create harm.
Why this matters
Shirali and Arul say public skepticism about AI is growing, and while healthy skepticism is appropriate, a blanket ban on potentially valuable tools would be regressive. They argue that increasing AI safety through rigorous, explainable testing is the way to build and restore public trust.
Those interested in learning more about Circuit Breaker Labs and other vetted startups can see them at the Startup Battlefield competition in downtown San Francisco, October 13–15, 2026.
