Forum AI

Today, we're launching Forum AI. Our mission: bring expert judgment to AI systems. We've assembled world-leading experts to evaluate how major AI models handle high-stakes questions—and create the new content these systems need for complex topics. Learn more: byforum.com

Worth a read. From Robbie Goldfarb on three capabilities that will define the AI systems we can truly rely on. forumai.substack.com/p/toward-wiser…

Please welcome Forum AI, our latest announced investment. @TheForumAI is building a platform that brings human judgement to artificial intelligence. medium.com/lerer-hippeau-…

And thank you to everyone for supporting us for these many months. We have an awesome team! So proud of what we have built!!

Our mission: make AI trustworthy on questions where getting it wrong has real consequences. As AI becomes the primary source for information, we need infrastructure that ensures these systems can handle questions where judgment matters as much as facts. More: byforum.com

Second, our experts create original content that doesn't exist elsewhere: real-time analysis on breaking news (often within minutes), capturing context, adding to mainstream coverage, plus training data for scenarios where quality examples simply don't exist.

We do two things: First, evaluate how ChatGPT, Claude, Gemini, and other major models handle high-stakes questions—assessing tone, balance, context, and identifying missing perspectives and hidden bias on topics from inflation to geopolitical conflicts to mental health.

Forum AI has assembled world-leading experts—former Treasury Secretary Larry Summers, former Speaker Kevin McCarthy, Fareed Zakaria, Scott Jennings, Niall Ferguson, former Assistant CIA Director Rob Kee, journalist Salena Zito and dozens more across economics, foreign policy, healthcare, and education.

At Meta, I saw algorithms shape what billions believe. As a journalist, I learned news sources can often hide bias behind objectivity. AI systems could make these problems exponentially worse- not because they're biased one way, but because they're fundamentally opaque.

I want to share something I've been working on for the last few months. We're launching Forum AI to bring expert judgment to AI systems. The challenge: AI can't distinguish between expert analysis and internet noise, synthesizing everything into confident answers while stripping away attribution.