“It is hard to see how you can prevent the bad actors from using [the technology] for bad things,” the Toronto-based researcher told The New York Times. In a subsequent tweet, he said he had left the search giant not to criticize it, but so that he “could talk about the dangers of AI without considering how it impacts Google.” (The New York Times, The Logic)

Talking point: Hinton is one of the pioneers of the machine-learning techniques and models underpinning much of the AI that labs and companies are currently developing and deploying. The British-born researcher arrived at the University of Toronto’s computer science department in July 1987 after a stint across the border; he joined Google in March 2013 after it acquired a startup he spun out of the institution. (He’s also the chief scientific adviser at Toronto’s Vector Institute, a cornerstone of the federal AI strategy.) Hinton said he’s grown concerned that AI systems are surpassing human intelligence in some fields. He’s concerned about the technology’s effect on the spread of misinformation and employment. And he’s worried that competition between tech firms and countries to roll out more advanced systems will lead to a kind of AI arms race.