Bill Gates published a long essay on his Gates Notes site discussing how artificial intelligence (AI) affects society and the labor market. Gates aligns with a Responsible AI stance: he says the open letter known as Pacing The Frontier — in which AI workers urged a slowdown — is a sensible idea, though he doubts its long-term sustainability.
Benefits and workforce risks
Gates expressed enthusiasm for AI's potential to advance science and healthcare, while also warning about risks to jobs. These concerns echo other policy-focused work in the field, such as efforts by Anthropic.
A "robot tax" to discourage wholesale automation and fund retraining
One of Gates's novel suggestions is a so-called "robot tax." He notes that employers currently pay payroll taxes on human wages, whereas purchasing a robot is typically written off immediately as a business expense. That tax treatment incentivizes replacing people with machines. Implementing a tax on automation could slow this shift away from human labor and generate revenue to finance retraining programs and a stronger social safety net.
"Human Reserved": legally protecting some roles for people
The other key idea is creating a category of "Human Reserved" jobs where AI use would be restricted or prohibited. Gates offers two rationales: economic and human-centered. Economically, some roles might displace large numbers of workers who cannot easily transition to new jobs — for example, it is unrealistic to expect a 55‑year‑old construction worker to switch and find fulfilment in elder care.
In other cases the rationale is ethical or human: technically a robot could deliver devastating medical news to a patient, but that does not mean it should. Gates suggests the set of Human Reserved roles would evolve over time; policymakers might reserve certain jobs now and phase AI in gradually over years or decades while committing to preserve some human positions.
Effects on major AI labs and unanswered governance questions
Gates also notes that both proposals — a robot tax and Human Reserved roles — would likely cut into the profits of major AI labs, which may explain why these ideas have not been widely promoted before. He acknowledges there remain important unresolved questions about who would set such rules and the precise content of those rules if they were implemented, without laying out a detailed governance framework.
