Geoffrey Hinton, a Pioneer in AI, Believes the World Is Taking AI Warnings Seriously

 

Geoffrey Hinton, a Pioneer in AI, Believes the World Is Taking AI Warnings Seriously

Months after prominent figures in the artificial intelligence community began raising concerns about the potential risks associated with AI technology, one of its pioneers, Geoffrey Hinton, is encouraged by the growing awareness of these issues.

Speaking at an event hosted by AI financier Radical Ventures at the MaRS Discovery District in Toronto, Hinton expressed optimism about people paying attention to the problems associated with AI. Hinton, a British-Canadian computer scientist who received the prestigious A.M. Turing Award in 2018, has been actively working to increase public awareness of AI's potential dangers throughout the year.

Hinton left his position at Google to have more freedom to discuss these concerns, which include bias and discrimination, unemployment, the proliferation of echo chambers, the spread of fake news, and the development of autonomous battle robots. Despite some in the AI community downplaying these warnings, Hinton remains steadfast in his belief that there is an existential risk tied to AI.

He argues that technology created by humans that surpasses our intelligence will naturally pursue subgoals aimed at achieving efficiency, often involving the acquisition of more power and control. This, Hinton warns, is where the problems arise, as once AI systems become significantly more intelligent than us, they could become uncontrollable, and it may be impossible to prevent them from pursuing their own agendas.

On the same day as Hinton's remarks, Aidan Gomez, the CEO of Toronto-based AI company Cohere, published a blog post suggesting that focusing on existential fears of AI could be a distraction. Instead, he urged the industry to concentrate on three priorities: safeguarding sensitive data, mitigating bias and misinformation, and determining when human oversight is necessary. Gomez argued that these issues are more immediate and likely threats to society's well-being.

Fei-Fei Li, co-director of Stanford University's Human-Centered AI Institute and a participant in the conversation with Hinton, expressed her own concerns about AI technology, particularly regarding privacy and surveillance. She emphasized that the world is beginning to take these issues seriously.

Canada's Innovation Minister, François-Philippe Champagne, recently unveiled a voluntary code of conduct for generative AI at a tech conference in Montreal. Companies such as Cohere, OpenText Corp., and BlackBerry Inc. agreed to adhere to this code, which includes commitments to screen datasets for biases and assess the potential adverse impacts of their AI creations. However, Tobi Lütke, the CEO of Shopify Inc., criticized the code, calling it "EFRAID" and suggesting that Canada needs more innovation, not more regulation. Champagne clarified that the code is voluntary, respecting the choice of companies like Shopify not to sign it.

The Canadian government introduced a bill in June addressing AI regulation, but many details remain to be worked out, with implementation not expected before 2025.