Worrying about AI as an 'existential' threat distracts from immediate risks, Balsillie says

Former tech executive Jim Balsillie says artificial intelligence should be regulated, but thinking about it as an existential threat distracts from the immediate challenges posed by other new technologies.

Balsillie, who was co-CEO of Research in Motion (now known as BlackBerry Limited), said in an interview on CBC's The House airing Saturday that politicians do need to pay close attention to new developments in artificial intelligence.

"We like cars, but we don't like drunken drivers speeding in front of schools. So we need to regulate to get the benefits and attenuate the harms," he told host Catherine Cullen.

But in the face of immediate problems linked to information technology — such as widespread data gathering and mental health problems made worse by social media addiction — thinking about AI as an "existential" problem is a "distraction," he said.

"These harms are happening now and we have to be very careful not to be drawn or duped into going into existential possibilities that are unquantifiable, both in their degree and timing," said Balsillie, who founded the Centre for Digital Rights.

Representatives of 29 governments signed on this week to what's being called the Bletchley Declaration — an agreement to take steps to ensure AI is developed safely that also warns of the technology's potential to cause "catastrophic" harm.

The leadership of DeepMind, Google's AI company, has said it's important to start thinking now about how to regulate a superintelligent AI system.

Elon Musk, the head of social media platform X, warned this week of an existential threat posed by AI. Meta's own head of AI, meanwhile, claimed Google is pushing for more regulation in order to make it harder for rivals to start competing AI companies.

But Balsillie described the debate about existential threats as "gaslighting."

"Is this a legitimate concern, these existential risks? Or is it a tactic of a bird feigning a broken wing to take you away from the nest of near-term regulation?" he said.

Balsillie spoke to The House this week after appearing before a parliamentary committee studying Bill C-27, the government's proposed legislation on AI.

C-27 would make several changes to Canadian law. It would introduce an Artificial Intelligence and Data Act, which would implement some regulations on "high impact" artificial intelligence systems.

Industry Minister François-Philippe Champagne has said that "high impact" might describe an artificial intelligence program that could determine whether a person receives a job or a loan.

"We are all grappling with the tremendous power of artificial intelligence, which offers great possibilities as well as risks,' Champagne told the Commons industry committee in September. Champagne also announced several amendments to the bill.