Jaron Lanier: 'No AI, Just People' — 10 Arguments Still Resonate

For AI researchers and product leaders, Lanier's claim is not merely philosophical. If 'AI is just people,' model outputs are aggregations of human contributions, and every benchmark, deployment, and safety review must account for the humans whose data and labor shaped the system. That reframing pushes the field away from anthropomorphizing models and toward auditing training pipelines, attributing outputs, and designing consent-aware data infrastructure.

On September 13, 2026, a transcript of a StarTalk Special Edition episode featuring computer scientist Jaron Lanier began circulating on Hacker News and SingjuPost. The episode's title, "There Is No AI (It's Just People)," captures Lanier's central challenge to the AI industry: that artificial intelligence should be understood not as an autonomous creature or separate intelligence, but as an artifact of human collaboration. Host Neil deGrasse Tyson, co-host Gary O'Reilly, and comedian Negin Farsad pressed Lanier on the evolution of the internet, social media addiction, alternate technology business models, and what the episode description calls data dignity. The resulting transcript offers a compact, philosophical counterweight to the more common AI discourse that treats models as independent agents.

“

Host Neil deGrasse Tyson, co-host Gary O'Reilly, and comedian Negin Farsad pressed Lanier on the evolution of the internet, social media addiction, alternate technology business models, and what the episode description calls data dignity.

Lanier's social media critique is an important thread. He discussed his earlier book 10 Arguments for Deleting Your Social Media Accounts, noting that the arguments have only become more pointed over time. In a moment that is both humorous and revealing, Lanier said high schoolers are forced to read the book, including his own daughter, and that he jokes with them that they must have done something bad. This matters for AI because social media platforms are also the infrastructure through which enormous volumes of human-generated text, image, and interaction data are collected, labeled, and ultimately used to train machine learning systems. If the social media business model normalizes extraction, the same logic can carry into AI development, where data contributions are absorbed into models without clear attribution or compensation. Lanier's argument therefore recasts data as a labor and dignity question rather than merely a raw material.

The transcript also records Lanier's explanation of why he pursued virtual reality in the first place. He explicitly rejected the idea that people were unhappy with reality and needed a replacement. Instead, he offered two reasons: first, to explore "extraordinary weird ways" for people to connect artistically and socially; second, to create a "palette freshener" that would let a vivid alternative make ordinary reality newly extraordinary when the headset came off. He described placing a flower or a mineral in front of someone after a VR session in the 1980s so that they would see it as if for the first time. This perceptual framing parallels his view of AI: not as a replacement for human cognition but as a tool that can refocus attention, surface patterns, and enrich human capability without dehumanizing it.