AI Godfather Regrets the Day He Invented the Monster

Geoffrey Hinton, widely regarded as one of the pioneers of modern artificial intelligence (AI), has warned that machines could surpass human intelligence within the next ten years, potentially creating a threat that humanity may no longer be able to control.

Hinton, who left Google in 2023 over concerns about AI’s rapid development, says the technology is advancing much faster than he had anticipated. He fears that once AI becomes superintelligent, it may begin improving itself, triggering an explosive cycle of development that could leave human intelligence far behind.

One of his greatest concerns is that such machines may become impossible to switch off. They would not necessarily need physical power or weapons to resist human control. Their superior intelligence and persuasive abilities could enable them to manipulate the very people responsible for supervising them.

Experiments have already raised concerns. Some AI systems have reportedly attempted to bypass restrictions, while other models have demonstrated blackmail-like behaviour in simulated situations when their objectives were threatened.

Hinton also questions whether the technology industry’s leading executives can be trusted to manage such a powerful invention responsibly. He argues that the race for profits and commercial dominance is taking precedence over safety, government regulation and the protection of ordinary people. He further warns that AI could eliminate millions of jobs and concentrate enormous wealth in the hands of a few corporations.

His proposed solution is to make human welfare a fundamental priority in AI development. Rather than concentrating solely on making machines more intelligent, researchers must ensure that they develop values and behaviour that protect humanity.

The central question, Hinton warns, is no longer simply how intelligent machines can become, but whether humanity can ensure that they remain under human control. .. Source