THE CHOICES WE MAKE (BILL GATES)
Bill Gates argues that AI could become either an extraordinary equalizer or a major source of inequality and disruption. His central message is that governments, companies, and civil society must build a deliberate transition plan now—before rapid AI adoption outpaces society’s ability to manage its consequences.
Core argument
AI is advancing unusually quickly because it works through devices people already own, uses natural language, and can learn from existing information and training materials—unlike earlier technologies that required long adoption cycles.
Gates believes AI will eventually equal or exceed human performance across many cognitive tasks, while robotics will extend that disruption into physical work. He argues that this makes the transition fundamentally different from past shifts such as the move from agricultural to office work.
He says neither the benefits nor the harms can be postponed: they are arriving simultaneously, while public institutions are inadequately prepared.
Three major risks
Permanent job displacement: AI could eliminate substantial numbers of entry- and mid-level jobs in white-collar fields—including customer service, sales, software engineering, legal support, lending, data analysis, and patient triage—while improving robotics could also affect construction, hospitality, and other blue-collar work.
Gates is especially concerned about young workers, since entry-level jobs provide a pathway into careers. He warns that new AI-related jobs may demand advanced skills that displaced workers cannot quickly acquire.
Expanded capacity for harm: AI can lower the skill and cost required for fraud, deepfakes, surveillance, disinformation, cyberattacks, bioterrorism, and potentially autonomous weapons. Critical infrastructure—such as hospitals, financial systems, water utilities, power grids, and government-benefit systems—could be vulnerable.
He also raises longer-run concerns about advanced AI systems behaving in unintended ways, potentially becoming difficult for their creators to control.
Damage to children’s development and human relationships: Gates worries that always-available, agreeable AI companions could become addictive and discourage children from developing resilience, social skills, and relationships with other people. He cites research suggesting heavier, more emotionally personal chatbot use is associated with worse feelings among users, particularly those with smaller social networks.
AI could also weaken critical thinking if students use it as a substitute for intellectual effort. Gates emphasizes that critical thinking will be especially important when misinformation and deepfakes can be personalized at scale.
Potential benefits
Gates does not advocate simply stopping AI development. He argues that, if deployed intentionally and broadly, AI could accelerate innovation in clean energy, climate mitigation, food production, disease eradication, medicine, education, and public services.
In healthcare, he highlights AI’s potential to help hospitals identify emergencies, aid diagnosis, improve patient follow-up, explain test results, and extend scarce mental-health resources. He cites Viz.ai, which analyzes scans for strokes and other emergencies and is used in nearly 2,000 U.S. hospitals.
In lower-income countries, he sees agriculture as an especially promising early application: AI could provide farmers with localized weather, seed, soil, crop-disease, and livestock guidance that is currently unavailable to many of them.
AI could make government services easier to access by simplifying applications for health insurance, student aid, food assistance, and similar programs—reducing bureaucracy for families and administrative costs for government.
In education, he sees value in AI tools that support rather than replace “productive struggle”: explaining concepts, asking questions, and withholding answers when students need to reason through a problem themselves.
More broadly, he argues AI could give individuals, small businesses, people with disabilities, and entrepreneurs access to capabilities that currently require expensive experts or large staffs.
Gates’s proposed response
Create new national and international AI institutions. Gates says existing agencies are too fragmented to manage a technology that simultaneously affects work, security, taxation, elections, education, energy, public health, and more. He calls for domestic bodies that coordinate across agencies and a new international framework drawing lessons from nuclear inspections, aviation regulation, and ozone-protection agreements.
Make governance democratic and inclusive. He argues AI policy should not be left primarily to AI companies. Workers, policymakers, educators, healthcare professionals, local officials, parents, religious leaders, students, and community organizations should help decide how AI is used.
Design a stronger social safety net. Since AI-driven job losses could reduce payroll and income-tax revenue just as retraining and income support become more necessary, Gates says governments need systems ready to assist displaced workers, people whose pay or hours decline, and communities hit disproportionately hard.
Create “Human Reserved” jobs or activities. Gates proposes setting aside some work for people even when machines could do it technically or cheaply. The rationale could be economic—preserving meaningful employment—or ethical and human, such as ensuring that a human, not a robot, delivers devastating medical news or leads certain care roles.
Rebalance taxation between labor and capital. He renews his earlier proposal to tax AI tokens and robots in some form. His reasoning is that employers pay payroll taxes when they hire people but can often deduct investments in machines, creating a tax incentive to replace workers. Revenue from targeted AI/robot taxes could fund retraining and social supports, though he says the policy should avoid blocking clearly beneficial applications such as lower-cost medicine or education.
What Gates commits to
Gates says he will press lawmakers, global leaders, and AI companies to prioritize equity and governance in AI policy.
The Gates Foundation plans to use AI in health, agriculture, and education, including efforts to make AI models available in languages used in countries where the foundation operates.
His concluding challenge is to ensure that AI’s gains reach people without wealth, power, or privileged access—and to preserve human dignity, social connection, and economic security during the transition.
https://www.gatesnotes.com/a-turbulent-ai-era-and-critical-choices-to-make
