Artificial intelligence is becoming more powerful every year, but not everyone believes this rapid progress is completely safe. One of the people raising concerns is Geoffrey Hinton, often called the "Godfather of AI." Hinton is a British-Canadian computer scientist whose groundbreaking research helped make today's AI systems possible. He also received the 2024 Nobel Prize in Physics for his contributions to machine learning. Although he spent more than a decade working at Google, he left the company in 2023 so he could speak more openly about the possible risks of advanced AI.
In a recent interview on the One Decision podcast, Hinton warned that future AI systems could begin thinking in ways that humans may no longer understand. At present, many large language models process information in ways that researchers can partly examine and monitor. However, Hinton believes future AI could develop its own internal methods of reasoning or even create its own language to communicate with other AI systems. If that happens, humans may not know what these systems are discussing or how they are making decisions.
According to Hinton, this possibility becomes especially worrying if AI grows more intelligent than humans. He explained that machines could eventually solve problems, make plans, and process information in ways far beyond our understanding. While this could lead to incredible scientific and technological breakthroughs, it could also create systems that humans can no longer supervise or control.
Hinton has previously warned that advanced AI might manipulate people, spread misinformation, or make decisions that are harmful to society. During the podcast, he said he would not be surprised if AI eventually developed its own way of thinking that people simply could not interpret.
