Campbell Brown’s Forum AI targets high-stakes AI evaluation

Campbell Brown, the former news chief at Meta, has launched Forum AI, a 17-month-old startup based in New York. The company raised $3 million last fall in a funding round led by venture capital firm Lerer Hippeau. Forum AI evaluates how foundation models—large AI models—perform on high-stakes topics such as geopolitics, finance, and hiring. Brown founded the company after realizing during her time at Meta that public chatbots would become the primary funnel for information despite being highly unreliable.

To build its benchmarks, the startup recruits human experts, including Niall Ferguson, Fareed Zakaria, former Secretary of State Tony Blinken, former House Speaker Kevin McCarthy, and former cybersecurity official Anne Neuberger. Forum AI uses these expert-designed benchmarks to train AI judges—automated evaluation systems. The startup aims to have these automated systems reach roughly 90% consensus with the human experts, a threshold Brown states the company has already been able to reach.

According to Brown, foundation model companies are extremely focused on coding and math, which leaves complex information topics prone to errors. Brown noted, “You hear from the leaders of the Big Tech companies, ‘This technology is going to change the world,’ ‘it’s going to put you out of work,’ ‘it’s going to cure cancer.’ But then to a normal person who’s just using a chatbot to ask basic questions, they’re still getting a lot of slop and wrong answers.” Beyond general inaccuracies, Forum AI’s evaluations of foundation models revealed a “left-leaning political bias” across nearly all models. Brown also cited Gemini “pulling from Chinese Communist Party websites” for stories that have nothing to do with China. She dismissed the current compliance landscape for AI as “a joke,” noting that existing standardized benchmarks and checkbox audits fail to catch critical edge cases.

Forum AI is betting its business on enterprise demand for liability management. While turning compliance interest into consistent revenue remains a challenge, Brown believes that businesses using AI for high-stakes decisions like credit, lending, and hiring will demand rigorous accuracy to manage their liability. This commercial pressure could force AI developers to prioritize getting answers right over simply optimizing for user engagement.

Why it matters

Forum AI is betting that enterprise demand for liability management—rather than just consumer interest—will force foundation model companies to prioritize accuracy over engagement. By using human experts to train automated evaluation systems, the startup aims to establish a commercial standard for AI reliability in high-stakes industries.