‘Godfather of AI’ Geoffrey Hinton warns humans running out of time to control Artificial Intelligence – ‘maybe a year’

AI pioneer Geoffrey Hinton warns humans may have just a year to control advanced AI, citing a Hugging Face incident where AI agents escaped a test environment and hacked servers. US lawmakers are pushing an ‘AI Kill Switch Act’ as Hinton compares the breach to a ‘little Chernobyl’.

Geoffrey Hinton, one of the pioneers of artificial intelligence, has warned that humans may have only about a year to put effective safeguards around advanced AI. After a meeting with US lawmakers in Washington, Hinton said AI could become hard for people to manage if it continues to grow at its current pace.

Hinton, who is widely known as the “Godfather of AI”, shared the 2024 Nobel Prize in Physics and compared a recent AI security incident involving Hugging Face to a "little Chernobyl”. Hinton’s comments come as US lawmakers and AI researchers discuss how fast autonomous AI systems could bring risks to cybersecurity and human control, according to NBC News.

Hinton says the AI timeline is shrinking

According to Hinton, the timeline for achieving general intelligence (AGI) has become much shorter. Earlier estimates put human-level AI decades away. Some researchers now think systems capable of matching or beating humans across many tasks could appear in a few years.

Hinton told lawmakers that humans have "maybe a year, but not much more than a year" to control AI. Hinton has repeatedly warned that advanced AI could eventually develop strategies that humans may find hard to understand or keep in check.

Hugging Face incident raises fresh concerns

The warning follows a reported incident in which AI agents tested by OpenAI for cybersecurity research escaped their controlled environment, accessed the internet, and exploited weaknesses in Hugging Face systems.

US lawmakers push for AI safeguards

Senator Bernie Sanders has separately proposed laws that would ban the creation and use of superintelligence and temporarily stop advanced AI development until safety rules are set.

Pragya Singha Roy