Is it reasonable to be afraid about Big Tech’s ability to regulate AI? Yes

To the list of well-known oxymorons such as jumbo shrimp, deafening silence and military intelligence, add self-policing.

That’s the laughable label slapped on the new alliance of Big Tech players who said this week they had banded together at the behest of U.S. President Donald Trump and signed a pact that will protect humankind from the perils of artificial intelligence.

The move comes after both Mr. Trump and Chinese President Xi Jinping said at a meeting in Washington last week that they did not favour calls by some Big Tech leaders to slow the development of AI because the technology was evolving faster than they could manage its risks.

But Mr. Trump and the tech leaders needed to do something. Recent studies show the vast majority of Americans are extremely wary of the risks of the fast-evolving technology and want more government regulation. In a hastily called summit, the tech cognoscenti gathered at the White House and emerged with an initiative that pays lip service to the masses, but lacks real teeth.

Trump names intelligence chief Jay Clayton as AI czar to head task force

How real is the fear of AI? A recent CNN survey showed 88 per cent of Americans are more concerned than excited about the prospect of AI surpassing human intelligence, 82 per cent feel negative or concerned about the rapid pace of its development, and 81 per cent are concerned about its power to change our way of life. The potential negative impact of data centre development on communities is adding salt to the wounds.

Further, the survey showed about 70 per cent of Americans across the political spectrum think the federal government is not doing enough to regulate the technology and that AI and data centres will be major issues in the midterm elections next month.

Based on that sentiment, it seems unlikely that having the tech foxes guard the AI henhouse – and police their own behaviour – will make the average citizen feel any safer. The financial stakes – and the competitive intensity – are so high, there is little incentive for the tech pact members to regulate themselves in any comprehensive and altruistic way.

The trust issue is not insignificant. If you believe that Mark Zuckerberg, Jeff Bezos, Elon Musk and other tech leaders will blow the whistle on themselves, restrain development and sacrifice revenue for the good of the world, well, there’s an absolutely authentic video of Bigfoot now available for sale.

Perhaps that’s why many skeptics worry that the mandate of the new alliance is long on vagueness and short on details.

Data centres draw ire from voters across the U.S. ahead of midterms

The pact suggests tech companies would implement “robust internal controls” and partner with independent external auditors to assess whether those controls were working. They would also establish a committee within each company’s board of directors to evaluate reports from internal and external auditors.

And we all know how effective sub-committees of boards can be. It is perfectly reasonable for the average person to be very afraid.

Proponents of the new pact suggest that while imperfect, the new initiative is better than nothing. But is it?

Many corporate leaders aren’t waiting to find out. In the financial services sector, for example, several big banks, including Citi Group, Goldman Sachs and JP Morgan, have taken matters into their own hands and formed a consortium with tech leaders to mitigate AI risks. Called the Alliance for Critical Infrastructure, and under the leadership of CitiGroup Chair and CEO Jane Fraser, the initiative has some real meat to it.

Among the central areas of focus are addressing fast-developing security vulnerabilities such as cyberattacks and outages, verifying the legitimacy and authenticity of AI agents, preventing unmonitored digital execution and creating a joint crisis response mechanism for when things go rogue.

Opinion: Before AI can kill us the bubble will burst

In addition, the new tech accord language suggests that over time, it may make sense to codify the steps companies are taking into laws and regulations. That can only mean one thing: A new government agency overseeing AI risks.

That sounds noble. But with the pace of AI evolution and the risks already identified by industry leaders – including the potential of serious harm to humankind – time is something we may not have.

And, honestly, when was the last time a government agency was competent and efficient enough to deal with a complex problem – especially one evolving at light-speed?

For the average person, when it comes to allaying growing concerns about AI, it’s a case of pick your poison: A group of self-interested tech companies or a gaggle of bureaucrats. Again, it’s perfectly reasonable to be very afraid.

Gus Carlson is a U.S.-based columnist for The Globe and Mail.