Guardians of the Truth: Campbell Brown’s War on AI Misinformation
Campbell Brown has spent her whole life chasing the truth. She started as a famous TV journalist and later became the first news chief at Facebook. Now, she sees history repeating itself as artificial intelligence changes how we get our information. This time, she is not just watching it happen. She started a new company called Forum AI to tackle the messy, complex subjects where AI often fails.
Forum AI looks at high-stakes topics like world politics, mental health, and finance. These are areas where there are rarely simple yes-or-no answers. Instead of relying on generic data, Brown’s team finds the world’s top experts to set benchmarks. They then train AI judges to evaluate how models perform based on those standards. For her politics work, she recruited big names like Fareed Zakaria and former House Speaker Kevin McCarthy. The goal is to get AI judges to agree with these human experts 90% of the time.
The Fight for Accuracy
Brown founded Forum AI about 17 months ago in New York. She remembers the exact moment she realized the world had a problem. It was right after ChatGPT first launched while she was still at Meta. She saw that all the world’s information would eventually flow through these AI funnels, but the output was often poor. She even felt a personal worry for her own children, fearing they would struggle to find the truth in a world of digital noise.
What bothered her most was that accuracy did not seem like a priority for big tech. Companies are very focused on making AI good at coding and math, but they often ignore the harder task of keeping information truthful. When Forum AI started testing models, they found worrying results. Some models pulled stories from questionable sources, while others showed clear political bias or missed important perspectives. Brown believes there are easy fixes to improve these outcomes, but the industry has a long way to go.
Lessons from the Past
Brown learned a lot during her years at Facebook. She watched what happens when a platform tries to optimize for the wrong thing. She admits they failed at many things they tried, and the fact-checking programs she built there no longer exist. The big lesson was that prioritizing user engagement over truth has been bad for society and has left many people less informed.
She hopes AI can break this cycle. Companies can either give people what they want to see, or they can give them what is honest and truthful. While it might sound dreamy to hope for an AI that only tells the truth, Brown thinks big businesses might be the ones to force the change. Companies using AI for things like insurance, credit, and hiring cannot afford to be wrong because of the legal risks. They need the tech to be right every single time.
Building a New Standard
Forum AI is betting its business on this need for accuracy. Brown thinks the current way companies audit AI is a joke. For example, when New York started requiring AI audits for hiring, more than half of the companies still had unchecked violations. Real evaluation requires experts who can find the edge cases and hidden problems that simple bots miss.
Last fall, Brown’s company raised $3 million in funding. She knows that people are skeptical of AI, and she thinks that doubt is justified. While Silicon Valley leaders talk about AI curing diseases or taking over jobs, regular people are still getting wrong answers to basic questions. Brown wants to close that gap and build a world where we can actually trust what the machines tell us.
