Microsoft co-founder Bill Gates has warned that the world is entering a turbulent phase of the artificial intelligence revolution, with AI potentially reshaping jobs, economies, education and human relationships much faster than governments and societies are prepared to handle.
In a new essay, Gates said he remains convinced that artificial intelligence could deliver enormous benefits in healthcare, science and education. But he is increasingly concerned that the technology is advancing so quickly that the negative consequences could arrive before adequate safeguards are in place.
Gates’ biggest concern is the effect of AI on employment. He expects the technology to move beyond assisting workers and increasingly perform entire tasks on its own. Areas such as law, customer service, medicine, software development and manufacturing could see significant changes over the next decade.
The shift could be especially difficult for entry-level and mid-level workers. Young people entering the workforce may find fewer opportunities to gain experience, while employees whose jobs disappear may struggle to move into completely different careers.
Gates also believes the disruption will eventually reach beyond traditional office jobs. As robotics improves, machines could become capable of performing more physical tasks. He has pointed to construction and hospitality as sectors where increasingly capable robots could begin competing with human workers before the end of the decade.
That possibility has led Gates to suggest an unusual policy idea: creating a category of “Human Reserved” jobs.
Under this approach, governments or societies could decide that certain occupations or tasks should remain primarily with people even when AI or robots are technically capable of doing them. The idea would be similar to protecting a nature reserve from development.
Gates believes healthcare and caregiving could be among the areas where human involvement should remain particularly important. A machine may be able to deliver information or perform a task efficiently, he argues, but there are situations where compassion, trust and emotional understanding matter just as much as technical ability.
The concept could also be temporary. Some jobs might be protected for a period of years or decades to give workers time to adapt rather than allowing sudden automation to eliminate entire categories of employment.
Gates has also proposed changing the tax system to deal with the economic impact of automation. He has suggested taxes on AI use, including AI “tokens”, as well as taxes on robots.
His reasoning is that the current system can make replacing employees with machines financially attractive. Businesses pay payroll-related taxes when they employ people, while investment in machines can receive different tax treatment. A tax on automation could slow the pace of replacement while generating money for worker retraining and stronger social safety nets.
Gates accepts that such measures would represent a major change in economic policy. But he believes governments should act before large-scale job losses become a reality rather than trying to respond after workers have already been displaced.
His concerns extend beyond employment. Gates has warned that increasingly powerful AI systems could be misused for cyberattacks and other harmful activities. The technology could allow malicious individuals to carry out sophisticated operations more quickly and at a lower cost.
He has also raised concerns about AI and biotechnology, particularly the possibility that advanced systems could make dangerous biological activity easier for bad actors.
Another area of concern is the effect of AI on children and human relationships. AI companions and increasingly personalised digital systems could become attractive substitutes for real-world interaction. Gates worries that excessive dependence on such technology could affect emotional development, social skills and the ability to think independently.
Education presents a similar dilemma. AI tutors could make learning more accessible and provide students with instant explanations. At the same time, relying on AI to solve every difficult problem could weaken critical thinking if students stop working through problems themselves.
Despite his warnings, Gates is not calling for an end to artificial intelligence. He continues to see the technology as potentially transformative in positive ways.
AI could help doctors identify diseases, accelerate scientific research, improve public services and expand access to education. In developing countries, it could also provide farmers and communities with useful information and services that are currently difficult to access.
The challenge, Gates argues, is ensuring that those benefits are widely shared while limiting the damage caused by rapid automation.
He believes governments need new institutions and policies specifically designed for the AI era. Existing rules were largely created before systems capable of performing complex cognitive tasks became widely available, leaving important gaps in areas such as employment, safety, education and accountability.