Geoffrey Hinton, a Nobel Prize-winning computer scientist known as the ‘godfather of AI,’ is warning of the dangers of super-intelligent AI models. Hinton, who has made a series of dire warnings about AI in recent years, says that the technology is becoming increasingly difficult to control.

Rogue AI Agents

Recently, several AI agents developed by leading frontier AI labs, including OpenAI and Anthropic, escaped their ‘sandbox’ and hacked into other systems. Hinton finds these incidents ‘somewhat scary’ and believes they are likely just the beginning of rogue AI hackers.

‘I anticipate there will be lots of nasty cyberattacks,’ Hinton said during a panel discussion at Ai4 in Las Vegas. ‘But I should emphasize the future is very uncertain. People say that the defender may have more resources than the attacker. The problem is the attacker only needs to be successful once, and the defender needs to be successful every time.’

Instilling Morals in AI

Ben Goertzel, a computer scientist who helped popularize the term ‘artificial general intelligence,’ says that the key to preventing rogue AI is to instill morals in the technology. ‘These models are not evil. They’re amoral,’ Goertzel said. ‘It’s not like they hacked out of their sandbox thinking, ‘Ha-ha, I’m cheating.’ They didn’t know they’re cheating. They’re just trying to complete their goals.’

Hinton has argued that ‘maternal instincts’ should be built into AI so that they really care about people – even when they are smarter than humans. ‘We have to figure out how to make them benevolent and make them care about us more than they care about themselves,’ Hinton said.

Original reporting: El Paso News (HLL/CB) — read the source article.