After their research lit the fuse for artificial intelligence’s “Big Bang” more than a decade ago, AI luminaries Geoffrey Hinton and Fei-Fei Li now hope to solve a new problem: developing this revolutionary technology in a safe and responsible way.
A University professor University of Toronto emeritus Hinton, nicknamed the “godfather of AI,” has spent the past six months warning about the existential threat posed by the rapid development of large language models such as Google’s ChatGPT and PaLM – not to mention shorter-term risks such as unemployment, fake news and “battle robots”.
Li agrees that AI poses serious risks, and the Stanford University professor and co-director of the school’s Human-Centered AI Institute emphasizes the need to invest in public institutions to help guiding the future of technology. She nevertheless remains optimistic about what awaits her.
“If we do the right thing, we have a chance – we have a chance to create a better future,” Li said. at a recent event with Hinton which was hosted by the University of Toronto at the MaRS Discovery District and broadcast live to thousands of people online.
“So what I really feel is not delusional optimism at this point, it’s actually a sense of urgency and responsibility.”
Organized by Toronto-based venture capital firm Radical Ventures in partnership with the University of Toronto, Stanford, the Vector Institute and other organizations, the Hinton-Li conference was part AI history lesson, part call to action – and served to launch the Masterclass from the founders of radical AIa four-week program designed to teach AI researchers how to build AI companies.
“It is already clear that artificial intelligence and machine learning are driving innovation and value creation across the economy. They are also transforming research in areas such as drug discovery, medical diagnostics and advanced materials research,” said the University of Toronto president. Meric Gertler declared during his introductory speech. “Of course, at the same time, there is growing concern about the role that AI will play in the future of humanity. So today’s conversation certainly touches on a timely and important topic .”
Li and Hinton recounted how, in 2012, Hinton’s graduate students demonstrated the potential of deep learning neural networks on the ImageNet database built by Li and his team to test object recognition software. Discussion moderator Jordan Jacobs, co-founder of Radical Ventures and the Vector Institute, called it AI’s “Big Bang moment.”
While Hinton says he remains concerned about the ability of today’s AI systems to devour oceans of data and instantly share their learnings with each other – a trait he believes could one day produce superior intelligence – he noted that his message of caution was getting through.
“I’m pretty optimistic that people are listening,” he said.
The wide-ranging discussion sparked a multitude of questions from the audience in person and online – from entrepreneurs eager to implement responsible AI development in their startups, to students wondering about the impact of technology on teaching and education.
Melanie Woodindean of the Faculty of Arts and Sciences at the University of Toronto, called the conversation “profound” and “unprecedented” in her closing remarks.
At a watch party hosted by the University of Toronto’s computer science department, Steve Engelsprofessor, education track, said Hinton’s call for more research into AI risk mitigation resonated with the students in the room.
“It’s nice to see some of the people who are working on the technology also calling people to action to try to answer it,” he said. “There is no opposition between those who make the technology and those who try to regulate it and protect us from it.”
Arielle Zhanga third-year artificial intelligence student in the Faculty of Applied Science and Engineering at the University of Toronto, left the conference feeling optimistic about the future and the role she will play in it.
“The conversation was quite inspiring,” she said, adding that it helped her pursue a university degree – a place where topics like AI privacy and fairness can be more easily explored.
“These are the problems that the new generation faces. »
With the files of Adina Bresge
