The Godfather of AI Warns AI Could Develop Its Own Unreadable Language and Think Beyond Human Understanding

Geoffrey Hinton, widely regarded as the "Godfather of AI," has issued a stark warning about the future of artificial intelligence, suggesting that AI systems could eventually develop their own internal language—unintelligible to humans. This concern comes shortly after Hinton was awarded the 2024 Nobel Prize in Physics for his pioneering contributions to machine learning. Speaking on the "One Decision" podcast, which aired on July 24, Hinton explained that while current AI models use English to perform "chain of thought" reasoning—allowing developers to trace their logic—this may not last. He cautioned that AI could evolve to create its own private language for communication and internal processing. "Now it gets more scary if they develop their own internal languages for talking to each other," he said. "I wouldn't be surprised if they developed their own language for thinking, and we have no idea what they're thinking." Hinton also expressed concern that AI might begin to generate harmful or "terrible" thoughts, raising questions about control and transparency. He noted that most experts believe AI will eventually surpass human intelligence, and when that happens, humans may lose the ability to understand or predict its actions. A longtime researcher at Google, Hinton has been one of the most vocal critics of AI's risks, warning that many tech leaders downplay dangers such as mass job displacement and the potential for AI to act against human interests. The only way to ensure AI remains safe, he said, is if researchers can develop systems that are "guaranteed benevolent." His remarks come amid a global surge in AI development. Tech companies are offering massive salaries to attract top AI talent, while governments are stepping in to shape the landscape. On July 23, the White House released an "AI Action Plan" calling for the relaxation of strict regulations in states with burdensome AI rules to accelerate innovation. The plan also emphasizes the need for faster deployment of AI data centers to support the growing demand.