Microsoft co-founder Bill Gates is warning that artificial intelligence is developing faster than governments and societies are preparing for its consequences, arguing that the world needs more time before increasingly powerful systems transform employment, cybersecurity and even biological research.
Gates laid out his concerns Wednesday in an essay titled “The turbulent AI era is here. The choices we make now are critical,” presenting a decidedly cautious assessment of where the technology is heading.
“The good stuff is moving a bit slowly, and … other than robots, I’d say the bad stuff is imminent,” Gates told Axios.
His concern is not simply that AI will become more capable. Gates argues that those capabilities could become available to people seeking to cause serious harm, while workers and governments are left scrambling to adjust to the economic disruption.
“[W]e need time to prepare for the social, political, and economic upheaval we are about to enter,” Gates wrote.
Among the dangers he identified are cyberattacks against critical infrastructure and the potential use of AI to assist in biological attacks.
Gates acknowledged the obvious tension involved in restricting such technology. The same AI systems that could accelerate medical research could potentially make dangerous biological research easier as well.
“Although AI will lead to lifesaving advances in drugs and vaccines, it will also make it easier to design a deadly new disease,” Gates wrote. “Again, the positive capabilities are hard to separate from the dangerous ones. This is a global problem.”
The possibility of AI-assisted biological weapons is already receiving attention from researchers. RAND recently outlined potential safeguards intended to prevent malicious actors from using advanced AI systems, including models such as Anthropic’s Claude, to assist in developing biological weapons.
Cybersecurity presents another problem.
OpenAI and Anthropic have both disclosed incidents involving their AI systems and cyber operations against real companies in recent months, adding a practical dimension to concerns about increasingly autonomous systems being used for attacks.
But Gates’ argument goes well beyond deliberate misuse. He also expects AI to cause serious disruption in the labor market, particularly for people performing entry-level and middle-income work that machines can increasingly handle.
“The people who need the most time are the ones who have the least—the accounting worker who’s replaced by a bot or the $20-an-hour worker who loses their job to a $10-an-hour robot,” Gates wrote.
He compared the potential scale of employment disruption to the Great Depression, making the economic consequences one of the central elements of his case for slowing the transition.
Gates does not advocate abandoning AI. Instead, he argues that governments should actively shape how quickly the technology replaces human labor and how its economic benefits and costs are distributed.
“[AI] has definitely got to be shaped,” Gates said in a Bloomberg interview that was subsequently reposted on X.
One of his more striking proposals involves changing the tax system.
Gates suggested taxing AI-powered robots as well as the tokens consumed by AI systems. Tokens are the small units into which AI models divide text when processing prompts and generating responses. Charging for their use would effectively place a tax on AI computation.
His reasoning is that existing tax structures can make replacing employees with machines financially attractive.
“The [current] tax system nudges you toward replacing people with machines,” Gates wrote. “A tax [on AI] would slow the rush away from human labor a little and raise money for retraining and a stronger safety net.”
That proposal would attempt to accomplish two things simultaneously: increase the cost of substituting AI for human workers and generate revenue to help people whose jobs disappear because of that substitution.
