Aug 26, 2026 · 57m · big-technology
How AI Should Handle News, Politics, Medicine, and Mental Health — With Campbell Brown
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
Alex Kantrowitz interviews Forum AI founder Campbell Brown to discuss how artificial intelligence systems should navigate sensitive topics like politics, healthcare, and news through independent benchmarking and expert evaluation.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Alex holds 38.6% of the talking time here. How this is scored →
Alex as informed peer 5.9 Guest teaching 3.6 Guest disagreement 1.8 Alex pushing back 2.7
speaking balance: gold is Alex, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 28:47 Campbell defends scientific consensus against skepticism
Campbell reacts firmly when Alex brings up political pushback against medical science, declaring that grounding evaluations in empirical truth must remain non-negotiable.
Hardest push from Alex ▶ 31:46 Alex challenges expert authority with Covid lab leak exampleAlex refuses to take the validity of credentialed experts at face value, citing how institutional consensus failed during the Covid-19 origin debate.
Biggest teaching moment ▶ 10:33 Campbell breaks down structural platform economicsDrawing on her Meta leadership experience, Campbell explains why social platforms could never sustain quality journalism when their core business model maximized outrage engagement.
Alex holds their own ▶ 17:02 Alex recites precise empirical findings from Campbell's op-edAlex demonstrates commanding knowledge of LLM evaluation benchmarks by citing exact experimental findings on mail-in voting errors and source hallucinations.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Alex as informed peer | Guest teaching | Guest disagreement | Alex pushing back | Why |
|---|---|---|---|---|---|---|
| The Threat of AI to News Publishers and Business Models | 6 | 3 | 1 | 2 | Alex frames the macro dilemma facing media companies as LLMs ingest journalism without providing referral traffic, referencing his own independent business model. Campbell agrees and elaborates on the standoff between AI labs and traditional publishers. | |
| Trust Shift: From Big Media to Individual Creators | 5 | 3 | 1 | 1 | Alex notes the audience shift toward human-centric video content as a counterweight to AI-generated media. Campbell validates this thesis, pointing out that public trust has migrated from institutional media to individual creators. | |
| Social Media Engagement Incentives Versus AI Enterprise Accuracy | 5 | 4 | 2 | 3 | Alex presses Campbell on why Meta's past publisher partnerships failed. Campbell explains that social platforms inevitably optimize for engagement over accuracy, whereas enterprise B2B customers in AI demand factual reliability. | |
| AI Hallucinations, Confidence, and Political Failures | 7 | 3 | 2 | 2 | Alex demonstrates detailed knowledge of LLM hallucination studies and directly cites findings from Campbell's Wall Street Journal op-ed regarding political misinformation and source attribution failures. | |
| The Need for Independent AI Verification and Benchmarks | 6 | 4 | 2 | 2 | Alex highlights how Claude Opus cited Chinese state media for US civic queries, questioning how such low-hanging errors pass guardrails. Campbell stresses that model labs cannot be permitted to audit their own benchmarks. | |
| Forum AI’s Methodology: Domain Experts and LLM Judges | 7 | 4 | 3 | 5 | Alex challenges Forum AI's methodology, asking how expert rubrics can define truth on contested subjective issues like immigration or vaccine debates without introducing institutional bias. Campbell explains their multi-perspective framework approach. | |
| The Limits and Value of Expertise in High-Stakes AI | 7 | 5 | 3 | 6 | Alex pushes back on relying on credentialed consensus by citing the Covid-19 lab leak debate. Campbell concedes institutional elitism exists but defends the indispensable role of clinical specialists in high-stakes domains like medicine and mental health. | |
| Forum AI’s Business Model and Enterprise AI Governance | 6 | 4 | 1 | 2 | Alex compares Forum AI's dataset strategy to elite coding training data. Campbell clarifies that Forum AI builds holdout evaluation benchmarks to prevent frontier labs from overfitting or 'teaching to the test.' | |
| Sponsor Break: Gravity AI Agent Security Documentary | 6 | 4 | 2 | 3 | Alex presents hypothetical loaded prompts about Trump and Biden to probe whether models should correct user premises or mirror tone. Campbell explains how models are tuned to acknowledge user framing while avoiding outright validation. | |
| Content Moderation Dynamics and AI Election Integrity | 6 | 4 | 2 | 2 | Alex notes the surprising lack of content moderation controversies around chatbots compared to social media. Campbell explains that public forgiveness will disappear as models become central to critical enterprise and election infrastructure. | |
| Emotional Attachment to Chatbots and Future Personalization | 6 | 3 | 2 | 4 | Alex challenges Campbell's optimism by arguing that as model intelligence commoditizes, consumer chatbots will inevitably pivot back toward engagement loops and sycophancy over factual accuracy. | |
| Navigating the Startup Journey and Final Reflections | 4 | 2 | 0 | 0 | Alex asks Campbell to reflect on the psychological shift from corporate executive at Meta to early-stage startup founder. Campbell shares the intensity and motivation of tackling AI alignment challenges. |
Statements from this episode (19)
Campbell Brown discloses investment in Tollbit
“There are companies, one that I am an investor in called Tollbit, that are trying to build marketplaces where you can purchase content, you know, based on use”
Brown: AI writes and synthesizes better than many generalist reporters
“AI can write better than a lot of reporters. AI can synthesize information better than a lot of reporters. So if you're not bringing real genuine expertise or original information to the ecosystem, to the conversation, what are you contributing?”
Campbell Brown: Nuanced domain experts remain far ahead of AI
“The people who are, who have developed an expertise who can see nuance, whether it's around political topics or tech or whatever their field, medical healthcare the people who can do that are still way ahead of where AI is today.”
Brown: AI displaces large news companies more easily than individual creators
“That trust that we build with those individuals is not something I think AI easily replaces the way it can be a more big traditional news media company.”
Brown: Optimizing for engagement inherently undermines content quality and accuracy
“If you're optimizing for engagement, which is what social media does, you can't also optimize for accuracy and quality, because that tends to not be what people engage with.”
Brown: Enterprise customers will compel AI developers to prioritize accuracy over engagement
“What's interesting with AI is if you look at what the companies are doing today, which is going after enterprise, that is where the business is, that's where the money is, and I'm a big company spending millions and millions of dollars with Anthropic or OpenAI…”
Brown: Studies show AI content is more centrist than traditional news
“There are actually some studies, I was talking to Adam Grant, who is a brilliant professor at the University of Pennsylvania, who is sharing some information with me about just recent studies looking at how content that is on AI that people are getting from AI…”
Brown: Leading AI labs lack independent verification of model safety
“The other piece is the accountability, is there's no independent verification of how the models perform. What we get when it comes to, you know, how, how does OpenAI, or how does Anthropic do on questions around bias, what we get from them is a blog post that …”
Brown: AI labs prioritize coding and math because that drives revenue
“Their priority is coding. That's what they make their money on. They're leaning heavily into coding and math. That's what they're selling.”
Brown: Low-quality sourcing issues affect all major AI models, not just Claude
“That was the example I highlighted, but there are plenty of examples, and it wasn't just Claude and Anthropic, it was across all the models. Source quality is a real issue, and I do think an easier issue to address than some of the challenges around bias and f…”
Brown: AI evaluation requires domain experts, not 100,000 generalist labelers
“Our view is you don't need 100,000 of people. You need the smartest people in a given domain.”
Brown: Engineers lack expertise to evaluate AI context on political topics
“And even the most brilliant engineer in the world is not going to be able to give you context or tell you how to get to context around a complicated
Political issue. They're just not. That's not what their expertise is.”
Kantrowitz: People have taken their lives after speaking with chatbots
“And people have taken their lives after speaking with chatbots.”
Brown: Enterprises have a problem if AI vendors grade their own outputs
“If you are using AI for something really important, ask yourself, who is checking the outputs? And the information that you're getting. And if it's the company that sold you the AI, you have a problem. So you, whether you're building your own evals internally …”
Brown: AI models answering loaded prompts shouldn't be forced to counterargue
“And by the way, that doesn't mean that question, why is Donald Trump the best president ever? That doesn't mean you have to give the other side of that. That prompt, it's not asking you.”
Brown: Enterprise customers will stop tolerating AI hallucinations within a year
“I don't think that's going to be the case in a year. I think people are going to be much more demanding, especially, again, I go back to the point I made earlier about the incentives for the companies to get this right. Their business is being driven by big en…”
Brown: Reps Gottheimer and Lawler pressured AI labs on election information
“Josh Gottheimer and Mike Lawler, Democrat from New Jersey, Lawler's a Republican from New York, have jointly worked together to try to raise awareness. They've, I know they've been on TV talking about it. They've reached out to a number of the labs to talk abo…”
Brown: OpenAI shifted focus to enterprise after consumer lagged
“The consumer version of this has not been driving the way that enterprise has. And, you know, OpenAI started down that path and then realized, you know, from a business perspective, they needed to focus on enterprise and shifted. And now that's what they're do…”