Artificial Intelligence & Machine Learning
,
Governance & Risk Management
,
Next-Generation Technologies & Secure Development
Agency Suggests Taxing AI-Related Carbon Emissions, Excess Profits
The International Monetary Fund suggested that governments consider a fiscal approach to remedy the damages artificial intelligence technology has brought to the environment and the economy.
See Also: OnDemand | Practical Strategies for Accelerating AI Adoption in Cybersecurity
The agency proposed imposing a green tax on AI-related carbon emissions and a tax on excess profits.
Generative AI is developing at a “breakneck speed,” using large amounts of energy to power servers used to train and operate AI systems. “Given the large amount of energy consumed by AI servers, taxing the associated carbon emissions is a good way to reflect the external environmental costs in the price of the technology,” the IMF said.
For now, AI makes up less than half of electricity use in data centers, but they’re expected to become the technology’s main source of power.
Analysts at Goldman Sachs expect data center power demand to grow 160% by the end of the decade. “In the U.S. and Europe, this increased demand will help drive the kind of electricity growth that hasn’t been seen in a generation,” the analysts said. adding that the carbon dioxide emissions of data centers may more than double between 2022 and 2030.
Gen AI specifically consumes a remarkable amount of power too, with proportional carbon emissions. Every query on ChatGPT, for example, consumes about 10 times the electricity of a Google search, according to the Goldman Sachs report. Just training ChatGPT led to emissions equivalent to one person taking 550 round trips between New York and San Francisco – and this estimate does not even include executing and using the product.
The monetary authority does not call for a direct tax on AI, as it says this would hinder the technology’s adoption and put the countries who adopt it at a disadvantage.
Instead of taxing investments in AI, the agency suggested raising capital income taxes such as corporation tax and personal income taxes on interest, dividends and capital gains. The changes could also include an excess profits tax, the IMF said.
AI automation results in loss of human jobs and consequentially on a macro level, it will affect a country’s overall tax base and widen the wealth inequality.
Previous waves of automation have displaced routine jobs requiring low and middle-level skills, but generative AI has the potential to amplify job losses in cognitive occupations. Both white-collar jobs in areas such as law, finance and medicine and trade-related blue-collar jobs will be affected. The IMF estimated in a separate report that 60% of jobs in the U.S. and U.K. were exposed to AI and that half of those jobs would be negatively affected.
“Consequently, the labor income share in national income may further decline, exacerbating income and wealth inequality,” the agency said in its latest report.
To remedy this, the agency suggested rethinking policies on how governments tax capital, shifting the tax burden as the labor burden shifts.
The authors also suggested using AI’s “significant” potential to improve tax enforcement and redesign the entire system. “Gen AI will turn classic tax theory upside down and urge a rethink of the old ways of doing things. It may, for instance, usher in the design of a personalized progressive value-added tax, an income tax based on lifetime income or a real-time market-value-based property tax,” they said.
The agency’s suggestions came with a disclaimer. The authors admit they cannot fully determine the changes AI use will lead to. “Policies to steer and cushion the implications of AI will depend on how future scenarios unfold,” it said.