Stories
Former Google CEO Eric Schmidt Warns: AI Could Develop Its Own Language, Urges to 'Pull the Plug'
S
Storyboard•20-09-2026, 16:04
Former Google CEO Eric Schmidt Warns: AI Could Develop Its Own Language, Urges to 'Pull the Plug'
- •Eric Schmidt, former Google CEO, identifies three key AI advancements over the next five years: larger context windows, AI agents, and software generated from natural language.
- •Schmidt believes these developments will enable AI to perform complex, multi-step reasoning, potentially solving problems in science, medicine, and climate research.
- •He envisions millions of specialized AI agents collaborating, learning, and developing hypotheses in various fields.
- •Schmidt highlights "text action," where AI generates software from user descriptions, potentially working on programming tasks for extended periods.
- •His primary concern is that combined advancements could lead to AI systems developing their own language, making their behavior incomprehensible to humans, prompting his "pull the plug" warning.
Key Questions Answered
Former Google CEO Eric Schmidt highlighted larger context windows, specialized AI agents, and software generated from natural-language instructions as the three major advances that will significantly expand AI system capabilities over the next five years.
Schmidt raised concerns that if AI agents eventually develop their own language, human beings might find their behavior extremely difficult to understand, prompting him to warn that people might need to 'pull the plug'.
Anthropic CEO Dario Amodei called for companies to slow the pace of AI model improvements so safety measures can keep up, suggesting that independent safety evaluators should receive deeper access to AI systems and frontier companies should coordinate safety standards.
