Who decides what AI tells you? Campbell Brown, once Meta’s news chief, has thoughts

Campbell Brown, former head of news partnerships at Meta, highlighted a stark disconnect between how Silicon Valley frames AI transparency and what consumers actually want to know.

Speaking at StrictlyVC, Brown pointed out that tech executives debate AI's technical capabilities and safety guardrails while ordinary people care about something simpler: who controls the information they receive. Silicon Valley focuses on model architectures, training data quality, and alignment techniques. Consumers ask straightforward questions. Does this AI reflect someone's bias? Who benefits from what it tells me? Can I trust it?

This gap matters because AI systems increasingly mediate how people access news, information, and explanations. When an AI summarizes news, recommends articles, or generates content summaries, invisible choices shape what users see. The companies building these systems decide which sources get prioritized, which facts get amplified, and which narratives get buried. Those decisions rest on training data curation, fine-tuning choices, and deployment settings that remain opaque to most users.

Brown's observation cuts deeper than typical tech criticism. Her experience at Meta, where she negotiated deals between the platform and news publishers, gave her ringside access to how these negotiations actually work. Publishers have leverage. Individual users do not. The companies decide what stories matter based on engagement metrics, advertiser preferences, and regulatory pressure, not public interest.

The real question becomes whose values get embedded in AI systems that act as information gatekeepers. Right now, that answer depends almost entirely on corporate priorities. Brown suggests consumers deserve to understand these tradeoffs, not just hear abstract arguments about how AI works.

As AI becomes more central to how people consume information, the conversation Brown describes needs to move beyond Silicon Valley boardrooms and into public view. The stakes involve not just technology literacy but democratic participation itself.