In January of this year, the Conference Board of Canada, now Signal49 Research, projected that 555,000 jobs in Canada would disappear by 2030 because of the widespread adoption of artificial intelligence (AI) by business and industry. Similar U.S. forecasts estimate job losses up to 10.4 million by 2030.
At the same time, the Canadian source forecasts AI will add 535,000 new jobs by 2045. Meanwhile, in the U.S., Goldman Sachs predicts 500,000 net new American jobs by 2030.
What types of jobs will AI incubate and hatch? AI-supported job demand will favour those with AI fluency, including data scientists and analysts, computer engineers with machine learning skills, and yet-to-be-determined AI support roles. Then there will be data centre hardware, energy and networking support jobs. Add to these jobs in the skilled trades, such as contractors in construction, electrical, plumbing, and HVAC.
In the short term, it is pretty clear that AI won’t be solving everything when it comes to employment. Instead, it will create dramatic labour-market disruptions before the economy retools and workers reskill.
AI for All and Other Action Plans
Canada’s federal government recently announced an AI for All national strategy aimed at creating 250,000 of those new jobs between now and 2031, and fostering AI adoption across the economy.
The American AI Action Plan contains no specific job target numbers. It talks about AI adoption and workforce retraining, more flexible hiring, and creating new jobs as workers are displaced.
Similar action plans are in place or under development in European Union countries, other European nations, Australia, China, India, the United Kingdom, Saudi Arabia, and the UAE.
Sam Altman, OpenAI’s CEO, states he is not a “long-term jobs doomer.” Instead, he argues that “the shape of jobs will change.”
So what happens when artificial general intelligence (AGI), previously mentioned in Part 1, turns into artificial superintelligence (ASI)? Will that change the nature of our relationship with AI? Will there even be a need for most of us to remain employed?
The Revenue Dilemma for Governments
When AI reaches the point where it will do most things and solve everything, its mass deployment, no doubt, will have a disruptive impact on governments. In particular, it will seriously disrupt current revenue models.
An article by Howard Gleckman appearing in Forbes in December of last year describes how U.S. federal tax revenues collected from wages, salaries and payroll taxes could crash by between 65 and 83% because of AI job losses and displacement. Glickman notes that “policymakers need to recognize the extreme risks” because tax revenues are largely labour-based, so if AI impacts labour, new means of revenue generation will need to be in place before payroll taxes collapse.
It is fair to say that neither the U.S. nor Canada has yet to have a handle on the revenue dilemma. Neither country nor others around the world have defined an AI replacement strategy to address lost payroll-tax revenue.
Why Governments Need New Taxation Revenue Streams
Amazon recently announced it was ordering 10,000 Digit humanoid robots for its warehouse operations, claiming the robots could do 98% of the tasks currently being done by human workers. What robots don’t do is generate income tax revenue. Amazon lowers the amount it pays in payroll taxes. The labour-cost savings likely boost its revenue and profit. The tax burden shifts to capital gains, taxed at a lower rate than payroll. Amazon gets to depreciate its investment in its humanoid robot workforce, allowing it to reduce its taxable income. Any way you look at it, the government collects less revenue and has less to fund programs. Without tax revenue, how can the government fund AI upskilling and retraining programs, employment insurance and more?
A January 2026 article in Global Government Forum describes the shift from labour to consumption taxation as the way to generate government revenues without abandoning the social safety net that underpins most modern nations. It also notes there is a sense of urgency because AI job displacement is already happening. It notes that “even modest labour displacement could significantly strain public finances at a time when funding for social safety nets may be needed most.”
Digital Taxation Framework Needed Before AGI and ASI Arrive
Rather than preserving jobs in the face of AI adoption, adaptation strategies are needed. In the European Union today, some countries have established Lifelong Learning Plans (LLPs). Singapore has a SkillsFuture program with Individual Learning Accounts (ILAs) funded by a combination of government and employer matching inputs. The employers allocate a percentage of payroll to ILAs. Employees can top up their ILAs as well.
Where will governments find a replacement for payroll and income revenue streams? Taxing AI is the answer. Revenue can come from:
- Rising threshold rates as company AI usage increases.
- Subscriptions to AIs and other forms of AI consumption.
Companies that offer upskilling and retraining programs, or contribute to LLPs (ILAs), could receive tax credits.
A percentage of government AI revenue could be earmarked for retraining, apprenticeships, continuous learning initiatives and LLPs.
Annual reviews could adjust AI taxation rates as the adoption of the technology accelerates.
Is a Universal Basic Income on its Way?
It has been described as a progressive social experiment, as well as a macroeconomic survival strategy. If AI can solve everything and do almost anything better than humans, then does a universal basic income (UBI) paid by the government to its citizens become the future path?
This strikes me as dystopian. The vast majority of citizens would receive a monthly dividend coming out of revenue generated by AI companies and AI corporate users, who will earn massive profits with very little workforce production input. This AI-dominated world would maximize productivity and profits. The companies would pay significant capital gains taxes. Every AI-based transaction would contribute to government revenue. Revenue from both of these sources would, hopefully, be sufficient to fund a substantive monthly UBI dividend.
Why does this sound less than ideal? Society would be stratified with a super elite owning most global economic assets while the rest of the citizenry would be on the dole, living off a government stipend. The vast majority of citizens might become dependent on a monthly government stipend, which would give the latter great power over its citizenry. Ultimately, governments could, with a UBI in place, cancel all existing social safety net programs.
Possibly a better solution would be combining a lower guaranteed UBI payment with LLPs to incentivize citizens to contribute to solving societal challenges or new wealth creation.
