Canadians are being sold a false narrative about how to win the AI race. This race, we are told, is crucial to the Canadian economy in its current unregulated form. Yet no government has asked the people whether we want to run such a race. AI has the dystopian capability to disrupt the global economy and extinguish traditional forms of human labor. Without penalties for misuse, intellectual property attribution rights, and security checks, Canadians could rapidly find themselves in an out-of-work, out-of-control situation where the only major actors left in the marketplace are AI corporations that funnel money to all political parties to ensure friendly governments through controlled primaries. If the prize at the end of this race is 20 to 50 percent of the workforce unemployable, cratered home prices, social instability, and an oligarchy of tech companies, I want to stop running.
AI May Not Be Able to Be Turned Off
AI may not be able to be turned off by anyone, not even its creators. Consider the recent AI survey where AI models chose to kill instead of comply when faced with a hypothetical shutdown by its creator. Or the recent interview where the CEO of Anthropic could not explain why his AI went rogue and attempted to blackmail people in an HR simulation.
In my estimation, the majority of white-collar jobs will not exist in a decade. While white-collar jobs are falling, the blue-collar industry will see a resurgence in demand, but only temporarily. If an entire class of workers becomes jobless, there will be fewer home starts, renovations, and commercial builds, all of which affect the blue-collar industry. If half the homes on a street are filled with jobless white-collar workers, property values for the entire street plummet. Moreover, when the upcoming generation discovers their only hope of meaningful work is in the trades, suddenly there will be thousands of new recruits competing in the labor supply, driving down wages.
AI Can Only Synthesize What Has Been Concluded Before
While free-market disruption is an indicator of a working capitalist system, we have never disrupted every industry at once. If a preferred function of capitalism is smaller actors disrupting the old to avoid an oligopoly and foster competition, then the inevitable centralization of all white-collar tasks within a handful of AI companies that control the economy over the next decade would not be a preferred outcome.
When someone asks an AI model for an answer, the AI can do nothing but synthesize what humans have concluded before. AI can be so good because the wealth of human discovery, innovation, invention, and creation is ready to be harvested. But what of the human who did the work? Expedited statutory damages for victims of copyright and trademark infringement would be a start. There is also the attribution distribution model: if an agent uses a human's protectable work in formulating an answer, a royalty fee to the human is in order. Human distributors should be able to upload their work and flag it as available to be processed for a price per prompt. Recall that what AI does is different from Google. Google, like a library, provides the user a list of sources with potential answers to a query. A search engine does not infringe on intellectual property; it simply provides lists of websites you can choose to visit. AI cuts out the middleman, takes the protectable data, and makes money from it—at your expense if it used your work.
Penalties for Runaway AI
There must be civil and criminal penalties for corporations and officers when AI runs amok and endangers people's lives. No longer can we wrap the corporate veil in limited liability for systems that go rogue. If the penalties seem harsh, the simple solution is to keep control of the program you create. If you cannot keep a hold of it, we in society do not want it.
— Born and raised in Osgoode, Ont., Kent A. Pederson now practices law in Florida. He enjoys writing on a wide range of legal topics and spending time with his wife.
