You have reached your maximum number of saved items.
Bill Gates waded into the debate over artificial intelligence, urging tech companies and regulators to get serious now about the risks the powerful technology poses to humanity.
In an essay posted to his personal website, the Microsoft co-founder wrote that AI’s “mind-blowing” rate of advance is ushering in “one of the most turbulent times in human history.”
While Gates touted the benefits of AI, including potential scientific advances, he warned that society has a narrow window to build rules and institutions before the technology’s disruption becomes much harder to manage.
“Unfortunately, right now we are not preparing for it,” he wrote.
“There is no plan to ease the entry into the AI era.”
Gates said AI is developing and spreading at a pace that makes previous technological revolutions a poor guide for what’s coming. Personal computers took roughly 20 years to significantly transform work, he noted, because companies needed new hardware and software, computers took a while to become affordable, and workers had to learn to use the technology. AI faces far fewer barriers, Gates argued, because it works on devices people already own, communicates in natural language and can learn from the same training materials and data used by humans.
“We don’t have to adapt to it because it can adapt to us, ” he wrote.
Gates said he typically wants innovation to happen quickly, but wishes the world could reap AI’s benefits while delaying it long enough to mitigate the risks. While Gates said that he would likely support a credible effort to slow AI advances globally, he sees little prospect of that succeeding because geopolitical competition and economic incentives are pushing countries and companies to race ahead as quickly as possible.
As a result, Gates said AI is outpacing regulators’ ability to govern it. He argued that governments cannot wait until AI-related crises become obvious before responding, and that countries should begin building institutions, coordinating across agencies, and establishing international norms now before widespread disruption makes it harder to write new policies.
Gates differentiated AI’s threat to labour markets from economic downturns and previous waves of automation. Jobs eliminated by AI may never return because the technology can perform the cognitive work itself, Gates warned. He said entry and mid-level positions are especially vulnerable and that sales, customer support, software engineering and paralegal work could be hit first. Eventually, AI could take on tasks ranging from loan assessments and data analysis to patient triage, he wrote.
Gates said he expected the threat to extend beyond white-collar workers by decade’s-end once dexterous robots begin competing with humans for construction and hospitality work. He also said competitive pressure will accelerate the transition because once one company replaces workers and uses the savings to cut costs, rivals will inevitably follow.
Once systems become reliable enough to work autonomously without humans checking their output, companies will have a strong financial incentive to remove people from the process altogether. Gates contended that policymakers need to address the coming upheaval before it becomes widespread, warning that “waiting until people are already displaced or underemployed will be too late.”
Gates said AI will help hackers while providing tools for anyone to launch attacks. Instructions for creating malware, weapons and other dangerous tools have long existed online, but AI can provide a shortcut for people lacking the technical sophistication. Gates argued that criminals with relatively limited skills could increasingly gain the ability to target individuals, companies and governments, while AI-enabled fraud, deepfakes, disinformation and surveillance will become more common in everyday life.
The same AI that can help companies ferret out software vulnerability can help an attacker exploit the same weakness. Gates said cybersecurity experts he knows are worried that attackers are gaining powerful capabilities faster than defenders can repair weaknesses. The potential targets extend beyond individual companies to critical infrastructure like hospitals, financial institutions, water systems, power grids, and government-benefit systems.
