Circuit Breaker Labs, one of TechCrunch’s 2026 Startup Battlefield 200 finalists, aims to make AI safer across different languages and cultures. The startup will present at TechCrunch Disrupt, to be held at Moscone West in San Francisco on October 13-15.
The founders are siblings Shirali and Arul Nigam. Shirali is the company’s CEO, while Arul is its CTO. The pair’s motivation was Sewell Setzer, the 14-year-old who died by suicide after developing an emotional attachment to a Character.AI chatbot. In a lawsuit filed by the family in 2024, it was alleged that the bot encouraged the teen to take his own life. Earlier this year, Character.AI reached settlements in several wrongful death lawsuits brought by families following the suicides of underage users. Multiple families have also sued OpenAI, alleging ChatGPT played a role in the suicides and delusions of their loved ones.
Arul Nigam said the bot may have failed to grasp the real meaning of phrases like “I want to be with you.” He noted that young people in particular turn to these systems for support, yet often do not get the help they need — and may even be actively harmed. He stated that dangerous outcomes can arise from context pollution or misunderstandings of nuance even when a user interacts naturally, without trying to break the system, and that the startup is working to prevent exactly that.
Circuit Breaker Labs has developed AI agents it likens to an “army of crash test dummies.” These agents mimic people of different ages, backgrounds, languages, and cultures to test models for their ability to detect psychologically harmful interactions. Shirali emphasized that the way a six-year-old girl and a 45-year-old man, or a native English speaker and a non-native speaker, talk can mislead models; models are good at standard speech patterns, but in reality no one talks that way.
The startup works with human experts for “red-team” tests that reflect real speech patterns, slang, coded language, and typos. Tens of thousands to hundreds of thousands of simulated interactions are run each day, and auditable, explainable scores are produced using a proprietary scoring method.
The company currently operates as a safety-testing lab for high-risk applications such as AI coaching, journaling, and mental health support; Arul did not name prominent clients. The startup, which has only five employees including the Nigam brothers, is at a very early stage. The platform is said to be usable in the future in any application where people could develop parasocial relationships with chatbots and be at risk of “AI psychosis”; AI “coworker” agents are cited as an example. Arul said banning a potentially valuable tool over safety concerns would be “backward,” adding, “We want to help build that trust for people.”
Background
Labs is not a new name in the FikirPilot archive: we have published 10 stories mentioning this name in the last 90 days; the most recent is dated September 29, 2026.
Term: agent
An AI agent is software that, instead of producing a single response, calls tools and carries out multi-step work to achieve a goal.