- Joshua Bengio is one of three AI “godfathers” who won the Turing Award for their breakthrough in 2018.
- He told the BBC that if he had known how fast AI was advancing, he would have prioritized safety.
- He called for regulations similar to those in the transportation sector and ethics training for computer scientists.
The professor, known as one of the three “godfathers” of AI, said: BBC He said he felt “lost” about his life’s work.
Yoshua Bengio added that had he known how fast the technology was advancing, he would have prioritized safety over utility, according to the broadcaster.
He earned the nickname “The Godfather” along with Jeffrey Hinton and Yann LeCun. won In 2018, it won the Turing Award, a prestigious computer science award, for breakthroughs in machine learning.
Concerns about AI have increased since ChatGPT went public last November. The main concern is that AI is getting closer to human-level intelligence than previously thought.
OpenAI CEO Sam Altman told Congress earlier this month that a government agency should be created to oversee AI and prevent it from growing out of control.
Bengio also told the BBC that regulation of AI companies and labs is needed.
“Governments need to track what they are doing,” he said. “This is the bare minimum we do for other areas, such as airplanes, cars, and pharmaceutical manufacturing.”
And according to the BBC, he asked for “ethics training” that computer scientists “normally don’t get.” “People who are close to these systems also need to get some kind of certification,” Bengio told the station.
he signed the contract on tuesday statement The AI Safety Center warns that the technology poses an “extinction” risk comparable to nuclear war.
In addition to Hinton, CEOs of top AI labs, including Altman and Google DeepMind’s Demis Hassabis, have also signed on.
Hinton quit his job at Google and said: new york times Earlier this month, he spoke about his regrets about pioneering AI. He said he was concerned that it would encourage misinformation and take away jobs.
The third “godfather”, LeCun, tweeted Many times he thought such beliefs went too far.
“Until we have a basic design for even dog-level AI (let alone human-level), it’s too early to discuss how to make it safe,” he said.