Why AI Is Repeating the Worst Mistakes of Social Media | Jaron Lanier

Summary

Computer scientist and futurist Jaron Lanier argues that AI companies are repeating social media’s central mistake: using systems that capture people’s attention and foster dependency as a business model. He says the concern is not simply the technology itself, but the incentives and design choices around it.

• AI and social media share an addictive dynamic. Lanier says conversational AI can be presented as an empathic, human-like companion—a pattern he compares to the 1960s ELIZA chatbot and to social media’s ability to manipulate attention. He warns that users may become emotionally reliant on chatbots.
• Value may be separated from its human sources. Lanier acknowledges that generative AI draws on human-created work, but argues that its outputs can make that collective contribution seem to come from the AI company instead. In his view, subscription payments often do not reach the people whose work helped make the systems useful.
• He distinguishes the models from the way they are deployed. Lanier describes large language models as potentially valuable tools for searching, extrapolating, and supporting collaboration. His objection is to turning them into “fake people” or autonomous agents that users are encouraged to defer to.
• He questions the case for AI agents. Lanier argues that many ordinary tasks—such as recipes, exercise, personal finances, and workplace productivity—do not require agents acting independently. He sees agentic AI as potentially useful in limited settings, including real-time ICU support or controlling drones, but says unnecessary autonomy can create risks without adding much value. He also cites scientific work involving proteins as an example of AI contributing without being treated as an independent agent.
• He proposes interactive tools that preserve human authorship. Rather than prompting AI for a finished result, Lanier suggests asking it to create an editor for exploring possible results. For example, a video-making tool might offer sliders for visual detail, historical references, or technical accuracy. He argues that experimenting with the options can build users’ understanding and intuition, keep them actively involved, and make their role as authors clearer.
• He is skeptical of promises of automatic abundance. Lanier questions the idea that today’s costs—such as data-center expansion and pressure on workers—will eventually be redeemed by a future of plenty. He points to companies increasing AI use without reducing employees’ hours as evidence that productivity gains do not automatically translate into more leisure or broadly shared benefits.
• He sees software and network effects as sources of power. Lanier says software increasingly shapes the actions available to people in areas such as commerce and public services, making it a powerful force in society. He also warns that network effects can reinforce natural monopolies and concentrate influence in a small number of technology leaders.
• He criticizes AI exceptionalism and calls for open debate. Lanier objects to describing AI as a new species or treating it as a godlike moral authority, arguing that this framing can encourage people to let systems operate unchecked. He also says technology workers should be able to criticize their companies without having to resign, and describes a lack of internal disagreement as a serious cultural problem.

Reviews, Guides, or Tutorials

None are presented. The discussion is an interview and analysis rather than a product review or instructional guide.

Main Speakers and Sources

• Jaron Lanier — Computer scientist, futurist, and Microsoft scientist; the principal interview guest and source of the analysis.
• Danny Fortson — Interviewer and US West Coast correspondent for The Sunday Times.
• Katie Prescott — Co-host of The Times Tech Podcast; joins the introduction and closing discussion.