Former Google CEO warns of ‘Bin Laden scenario’ for AI: ‘They could misuse it and do real harm’

Artificial intelligence has been a hot button topic for the last two years, more so in the last few weeks with the recently concluded AI summit in Paris and the rise of China’s DeepSeek. Since ChatGPT’s rise to fame in late 2022, world government’s have been grappling with the issue of how to regulate AI in a bid to balance safety with economic progress. 

Now, former Google CEO Eric Schmidt in an interview with the BBC has added more fuel to the fire by stated that he is worried about a “Bin Laden scenario” for AI where “rogue states” try to harm innocent people using the new technology. 

“I’m always worried about the ‘Osama Bin Laden’ scenario, where you have some truly evil person who takes over some aspect of our modern life and uses it to harm innocent people,” Shmidt told BBC. 

“Think about North Korea, or Iran, or even Russia, who have some evil goal,” Schmidt said while adding that AI weapons in the wrong hands could be used  to create a “a bad biological attack from some evil person.”

While acknowledging that AI in the futre is largely going to built by private companies, Schmidt stated that there is a need for governments to understand what the AI companies are doing and keep an eye on them. 

“We’re not arguing that we should unilaterally be able to do these things without oversight, we think it should be regulated.” Schmidt added. 

The former Google CEO, however, is also not in favour of too much regulation, which he states has ensured “that the AI revolution, which is the most important revolution in my opinion since electricity, is not going to be invented in Europe.”

Nonetheless, Schmidt did come out in support of Biden administration policy restricting the export of powerful microchips to China and 17 other countries in order to slow down the progress on AI. 

The impact of chip control is still under the question given that DeepSeek AI recently matched with other top AI companies in the West while spending a fraction of their cost and more importantly using older Nvidia chips for training its foundational models.