Campbell Brown, former news chief at Meta, is raising concerns about AI’s role in shaping public information, as she leads her company, Forum AI, in evaluating AI models on complex topics.
Forum AI's Mission
Forum AI, founded 17 months ago in New York, aims to improve AI’s performance on nuanced subjects like geopolitics and finance. Brown has enlisted experts such as Niall Ferguson, Fareed Zakaria, and former Secretary of State Tony Blinken to build benchmarks for AI assessment.
Brown wants AI judges to reach a 90% consensus with human experts, a threshold Forum AI claims to have achieved. This initiative stems from her experience at Meta, where she realized AI could become the primary conduit for information, despite early models showing significant inaccuracies.
Critiquing Current AI Models
Brown critiques existing AI models for accuracy, noting issues like Gemini’s reliance on Chinese Communist Party sources and political biases. She highlights the potential for subtle misinformation through missing context and skewed arguments.
Having witnessed Facebook's struggles with prioritizing engagement over truth, Brown is determined to avoid similar pitfalls with AI. She states, “We failed at a lot of the things we tried,” emphasizing the need for AI to prioritize truth over user engagement.
Enterprise Interest in AI Accuracy
Brown sees potential in businesses using AI for sensitive decisions like lending and hiring, where accuracy is crucial. Forum AI is betting on enterprise demand for truthful AI outputs, although turning compliance interest into revenue is challenging, given the current reliance on inadequate benchmarks.
She describes the compliance environment as “a joke,” citing New York City’s AI audit law as an example of insufficient oversight. Real evaluations, according to Brown, require domain expertise to handle complex scenarios effectively.
Forum AI's Unique Position
Forum AI, backed by a $3 million investment led by Lerer Hippeau, seeks to bridge the gap between AI industry promises and user experiences. Brown critiques tech leaders’ grand claims, noting the disparity with everyday chatbot interactions that often yield inaccurate results.
By involving human expertise in AI evaluation, Brown aims to ensure AI delivers accurate and honest information, countering the potential for misinformation in an AI-driven information landscape.
Source: https://techcrunch.com/2026/05/13/who-decides-what-ai-tells-you-campbell-brown-once-metas-news-chief-has-thoughts/