The Case for Boycotting Generative AI

Substack · Watch/Read · (gathered 2026-09-10)

• Gary Marcus advocates for a public boycott of generative AI, asserting that frontier labs are knowingly developing dangerous, irreversible systems without public consent, and both industry and government lack serious plans to mitigate the escalating risks.
• He cites candid admissions from Anthropic and OpenAI employees regarding significant risks, including a >10% chance of AI killing all humans and an unsolved “alignment problem,” leading Marcus to fear a high probability of “dystopia” from AI-generated pathogens, disinformation, and infrastructure hacks.
• Marcus identifies the core issue as the unreliability of “incorrigible AI” built on “probabilistic large language models,” which are “reckless and difficult to control,” contrasting them with stable “deterministic AI” and arguing there’s no commercial justification for pursuing such unstable systems.
• He urges consumers to boycott generative AI tools to apply economic pressure, forcing companies to pause development of untrustworthy systems and find reliable solutions, while also suggesting a global deal with China on AI regulation is a viable path forward.