“My fear is that humanity will completely lose control because we are creating beings more powerful than ourselves,” he admits Newspaper Quebecer Mati Roy, who holds a key position at the Californian company OpenAI, which developed the ChatGPT conversation robot.
• Also read: We tried Gemini for you
• Also read: Before you meet up with an AI virtual girlfriend, read this
• Also read: Artificial intelligence: Sora video generator raises fears about US elections
“Because nuclear energy can be used for good or evil, there is a risk of accident if artificial intelligence creates an extremely powerful being without aligning it with our values. “It could be extremely dangerous for humanity,” whispers the man in his thirties from Longueuil calmly on the other end of the line.
- Listen to the interview with Mati Roy, strategic managers of the OpenAI technical program, on Yasmine Abdelfadel's show QUB :
After studying engineering physics at the University of Laval, Mati Roy became interested in the effects of artificial intelligence (AI). He quickly rose through the ranks and joined the elite OpenAI team further developing the powerful ChatGPT tool.
For two years, he has been one of the strategic managers of the technical program at OpenAI, a company co-founded by Sam Altman and Elon Musk that would be worth around $108 billion, according to the New York Times.
Mati Roy is one of OpenAI's 1000 employees. Photo provided by Mati Roy
“I supervise people who claim to be ChatGPT to see how ChatGPT should learn. My job is to ensure that this data is of high quality,” he explains.
“We want to determine the ideal behavior of ChatGPT,” continues the man, who lives in the US state of Georgia.
Sam Altman is 34 years old. The Chicago businessman now runs OpenAI, which some are calling “the next Google.” Photo AFP
Battle between good and evil
A year ago, Quebec artificial intelligence expert Yoshua Bengio said he feared the excesses of a dictator if we didn't urgently regulate AI.
Yoshua Bengio testified before a US Senate subcommittee about the dangers of artificial intelligence. Photo of Yoshua Bengio's appearance before a U.S. Senate subcommittee
“All it takes is one election and the election of a dictator to take all means and control,” he warned. A few months later, Yoshua Bengio went so far as to voice his concerns at a US Senate subcommittee in Washington.
Earlier this month, the Quebec Innovation Council (CIQ) called on the Legault government to modernize labor laws to prevent advances in artificial intelligence (AI) from directly hitting the labor market.
Eternal life
When we ask Mati Roy about the risks associated with AI, he gives a similar critical speech to Professor Bengio, but is also quick to point out that well-managed AI with good values could contribute to the advancement of humanity. right rail.
Certainly, the creation of multiple artificial intelligences more powerful than humans is “very dangerous to the survival of humanity,” he emphasizes.
However, Mati Roy is convinced that humanity will take a big step forward if we manage to instill values in the machine to prevent our destruction.
“I want people to be able to live as long as they want,” says the manager.
“If we create creatures more intelligent than us, technology will advance very quickly. Problems like cancer become easier [à régler]», he concludes.
Can you share information about this story?
Write to us or call us directly at 1 800-63SCOOP.