Canada needs to move “as quickly as possible” in passing a proposed AI regulation bill, Yoshua Bengio, one of the godfathers of AI, told the National Post as international experts warned Tuesday artificial intelligence is an existential threat.
“Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war,” reads the one-line statement signed by industry leaders, including the CEOs of OpenAI and Google Deepmind, and international AI experts, including Bengio.
A year ago the Liberal government introduced a privacy bill, C-27, that includes a section focusing on artificial intelligence systems. The Artificial Intelligence and Data Act (AIDA) is aimed at “high-impact” AI systems and would ensure they “meet the same expectations with respect to safety and human rights to which Canadians are accustomed,” according to the government. The bill would also “prohibit reckless and malicious uses of AI” that cause serious harm.
Some have criticized the government for moving slowly on Bill C-27 while prioritizing other internet-regulation legislation, like the online streaming bill. It took nine months for the legislation to make it to second reading, after which bills move to committee for study. That study still hasn’t begun for C-27 though the bill completed second reading a month ago.
Since Bill C-27 was introduced last year, the landscape has changed with the launch of generative-AI systems like OpenAI’s ChatGPT, and in recent months experts have called for a moratorium on AI developments.
Governments take a long time to build regulation and legislation, he pointed out. “I’m still concerned that as usual, things will move very slowly. And for example, the next U.S. election is just around the corner.”
As a first step, Canada should move to pass AIDA “as quickly as possible” and then move for it to come into force faster than the two years the government has projected it will take. He said Canada should also “play a leadership role on the international front, to start the discussion about guidelines, eventually treaties, regarding the minimization of all kinds of AI risks.”
He said governments should be thinking about putting in place “very simple, short term” rules, such as penalties for impersonating of a human, and making sure that humans or corporations are behind social media accounts.
Governments need to take action quickly, within the next year, Bengio said.
Source: National Post