Microsoft’s Bill Gates Says AI Could Be ‘Most Dangerous Thing’ Humans Have Ever Faced, Warns Regulation Is Lagging

AI’s rapid evolution is intensifying debate over safety and oversight.

• Gates says AI has already reached dangerous thresholds in cyberattacks and biological threats.
• Tech leaders remain divided over how to balance AI’s potential with risks from increasingly capable systems.
• Microsoft’s AI gains have lifted its business, but investor concerns over massive infrastructure spending continue to weigh on the stock.

Microsoft Corp. (MSFT) co-founder Bill Gates has issued one of his starkest warnings yet about artificial intelligence, saying that the technology has already crossed dangerous thresholds in cyberattacks and biological weapons — while governments remain unprepared to control what could become one of humanity’s most consequential technologies.

In an interview with New York Times columnist Ezra Klein, Gates said AI is now “the most dangerous thing that humans have ever gone near,” arguing that existing regulatory structures are nowhere close to what the technology requires.

The comments mark a significant shift from Gates’ more measured assessment of AI risks in 2023, when he described them as manageable.

Gates Warns AI Has Already Crossed Cyber, Bio Thresholds

Gates distinguished between near-term risks from humans using increasingly capable AI systems and a longer-term scenario in which AI itself could become difficult to control.

He said AI can already give malicious actors capabilities that previously required highly specialized expertise.

“A.I. has crossed the threshold that its ability to empower a bioterrorist to kill hundreds of millions — that exists today,” Gates said. “The ability to do a cyberattack that scrambles all of the bank accounts, shuts down the electric grid — that exists today.”

He argued that these are no longer hypothetical risks, saying, “we crossed the cyber threshold and we crossed the bio threshold early this year.”

Gates also rejected the idea that technology companies can be trusted to police themselves, particularly as open-source models become more capable.

“You can’t rely on the industry to self-regulate here,” he said, arguing that government-enforced supervisory layers should apply even to open-source AI systems.

His longer-term concern is more existential, describing it as an AI system that becomes substantially more capable than humans and operates beyond effective human oversight.

AI Leaders Split On Doomsday Debate

Gates’ warnings come as the AI industry itself remains divided over how seriously to treat existential risks.

Microsoft CEO Satya Nadella has said any pursuit of superintelligence must remain grounded in the principle that AI should help humanity and remain under human control.  

OpenAI CEO Sam Altman has similarly warned that AI could go badly if humans lose control of increasingly powerful systems, while also cautioning against concentrating extraordinary AI power in a single person, company or country.  

SpaceX and Tesla founder Elon Musk has long described AI as an existential risk while simultaneously pushing aggressively toward increasingly capable systems. Recent reporting places Musk among technology leaders calling for greater caution as AI capabilities accelerate.  

Nvidia CEO Jensen Huang, meanwhile, has emphasized AI’s accelerating economic and technological benefits, describing AI infrastructure as productive infrastructure and arguing that AI is already creating real value.  

President Donald Trump’s administration has likewise prioritized accelerating U.S. AI development and maintaining American technological leadership, while pairing that push with national-security safeguards. 

On Tuesday, Trump hosted top AI executives at the White House, including Anthropic’s Dario Amodei, OpenAI’s Greg Brockman and Meta’s Mark Zuckerberg, where companies signed a voluntary AI accord pledging internal controls and independent audits. 

Microsoft’s Muted Year

Gates stepped down as Microsoft’s CEO in January 2000, handing the top job to Steve Ballmer while remaining chairman. He stepped down as Microsoft chairman in February 2014, ending his formal leadership role at the company while remaining a board member and technology adviser.

Microsoft gained tremendously from its early investment in and partnership with OpenAI, which helped bring the latter’s AI technology to Microsoft’s cloud and software services before rivals caught on to the trend. 

Late last year, Microsoft and OpenAI significantly reworked their partnership as OpenAI transitioned into a public benefit corporation. Under the October 2025 deal, Microsoft retained roughly a 27% stake in OpenAI and maintained key intellectual-property rights through 2032. 

Even though cloud and AI-led gains have translated into higher sales and profits, MSFT stock has remained subdued lately, rising a mere 6% year to date, as more investors grow concerned about the returns from the massive investments in AI development and data centers.

On Stocktwits, the message volume for MSFT has dropped 40% in the past year while the watcher count for the stock increased 6.4%. On Wednesday, the retail sentiment for the stock was ‘bullish.’

For updates and corrections, email newsroom[at]stocktwits[dot]com.<