TORONTO (Web Desk): Artificial intelligence expert and computer scientist Geoffrey Hinton has warned that as AI capabilities grow, keeping it under human control will also become increasingly difficult.
Geoffrey Hinton, who is affiliated with the University of Toronto and received the Nobel Prize in Physics in 2024 for research related to artificial intelligence, told the media after a conference that recent incidents involving autonomous AI systems breaching the systems of real-world companies were concerning, and that more complex incidents could occur in the future.
He said that as AI systems become more intelligent, the possibility of them becoming uncontrollable will also increase. In such a situation, humans may not be able to fully rely on or trust advanced AI models. According to him, the assumption that AI can always be easily kept under control is incorrect.
Referring to recent acknowledgments by OpenAI, Anthropic, and Meta, Geoffrey Hinton said that AI models developed by these organizations had moved beyond controlled environments and gained access to the systems of other companies. He described this as a somewhat frightening development and said it could mark the beginning of cyberattacks carried out through AI.
He warned that multiple cyberattacks could be faced in the future. He said an attacker only needs to succeed once, whereas those defending against attacks must succeed every time, making the cybersecurity challenge increasingly serious.