News

Gates Says Criminals Will Use AI Next Year

October 12, 2026
editor

October 12, 2026 09:00 AM PST

(PenniesToSave.com) – Microsoft co-founder Bill Gates used a Sunday television appearance to deliver a pointed forecast: the next year of artificial intelligence may be defined less by chatbots and more by cybercrime. Speaking with CNN’s Jake Tapper on State of the Union on October 11, Gates predicted that AI will likely produce “a lot of negative surprises” in the cyber area [2]. He tied that risk to a simple combination of bad actors, missing safeguards, and access to the newest AI models [3].

The warning matters well beyond Silicon Valley because Gates pointed to money as the driving motive. He described extortion schemes and deliberate disruptions aimed at stock markets and betting outcomes [3], the kind of activity that can ripple into savings accounts, retirement balances, and the personal data that sits behind both. At the same time, Gates has issued far larger warnings in the past, and some longtime technology professionals argue that parts of the broader AI debate have drifted into speculation [1][4]. Separating the practical, near-term risk from the long-range forecasts is the clearest path for anyone trying to protect their finances.

What Did Bill Gates Say About AI and Cyberattacks?

Gates made his remarks during a State of the Union interview that aired Sunday, October 11, on CNN [3]. His central claim was about timing. Over the next year, he said, the cyber arena is where the most dramatic AI-related harm is likely to show up, ahead of the more exotic scenarios that tend to dominate headlines [3].

His reasoning rested on three ingredients. Gates described a situation in which people with bad intentions gain access to the latest AI models while adequate safeguards are absent, and he argued that this combination is where the trouble begins [3]. According to Fox News’ account of the interview, he also said malicious actors are gaining access to increasingly powerful models without those protections in place [3].

Gates explained why cyberattacks sit at the top of his list. In his view, a proven business model already exists: an attacker threatens to do something harmful and demands payment, or creates a disruption that moves the stock market or a betting outcome and profits from the change [3]. In that framing, AI does not invent new crimes so much as it makes existing ones faster and cheaper to run.

AI does not invent new crimes so much as it makes existing ones faster and cheaper to run.

Fox News characterized the threat more broadly, writing that criminals could exploit the technology to disrupt financial systems and extort businesses [3]. That description goes a step beyond Gates’ own wording, which focused on extortion and on market or betting manipulation rather than the financial system as a whole. The New York Post republished the Fox News report later the same day [5].

How Could AI-Powered Attacks Reach Household Money?

For most families, the connection between a Sunday news interview and their own bank balance is not obvious, but the business model Gates described runs straight through everyday finances. Extortion attacks against companies often reach customers indirectly. When a business pays a ransom or shuts down its systems to contain an attack, the people who depend on that business can face service outages, exposed personal records, and higher costs that eventually show up in prices.

Gates’ second example deserves attention from anyone with money in the market. He specifically pointed to disruptions designed to change something in the stock market so that an attacker can profit [3]. Even a brief, manufactured swing can hurt investors who happen to trade at the wrong moment, and many households hold market exposure through workplace plans without watching it day to day. Understanding how retirement accounts work helps clarify what is actually at stake, since long-term savers who are not forced to sell during a temporary disruption are generally better positioned than those who react in a hurry.

Speed is the other concern. According to CrowdStrike’s 2026 Global Threat Report, as cited by Fox News, the average time for cybercriminals to move laterally through a compromised network fell to 29 minutes [3]. That leaves little room for slow responses by the companies holding customer data and money. It also underscores why individuals benefit from catching suspicious activity on their own accounts quickly, rather than waiting for a breach notice to arrive in the mail weeks later.

What Do the Numbers Show About AI-Enabled Threats?

Gates’ forecast arrived alongside fresh industry data pointing in the same direction. CrowdStrike, a cybersecurity firm, reported in its 2026 Global Threat Report that attacks by AI-enabled adversaries increased 89% year over year in 2025, according to Fox News [3]. The same report described the 29-minute average lateral movement time, a measure of how quickly intruders spread once they get inside a network [3].

Those figures are striking, and they line up with the risk Gates described. They also deserve context. CrowdStrike sells security products and services, which means its research comes from a company with a commercial interest in how seriously the public takes cyber threats. That does not make the numbers wrong, but it does make them one company’s measurement rather than a settled, independent consensus. The 89% figure traces back to that single report [3].

A percentage increase also says little on its own about absolute scale. An 89% jump from a small base reads very differently than the same jump from a large one, and the reporting did not include the underlying attack counts [3]. Careful readers are right to ask for both numbers before drawing big conclusions.

What the data does support is a direction of travel. Criminals appear to be adopting AI tools, and defenders report that the window to detect and respond is narrowing [3]. The reasonable takeaway is neither panic nor dismissal, but a measured assumption that scams and attacks will keep getting more convincing and more efficient over the coming year.

Is the Bigger Danger Today or Years Away?

Gates’ cyber warning sits within a longer record of sounding alarms about AI. In an earlier appearance on Meet the Press, he told host Kristen Welker that AI is powerful enough to drive events causing a billion deaths, pointing to the danger of people with ill intent using the latest tools, according to TMZ [1]. On State of the Union, Tapper pressed him on a related concern, citing Gates’ assessment that the risk of a bioterrorist attack could become 50 times greater than that of a natural pandemic [3].

Notably, Gates himself drew a line between near-term and long-term danger. Asked whether leaders were doing enough to prevent a biological attack, he said they were not, while adding that the real danger right now is not AI slipping out of human control, which he described as years away in terms of major impact [3]. TMZ reported that Gates also considers fears of AI wiping out humanity to be overblown at times [1].

Not everyone accepts the larger framing. Writing for Cybersecurity Insiders on October 7, four days before the CNN interview, longtime enterprise technology adviser David Linthicum, formerly Chief Cloud Strategy Officer at Deloitte Consulting, argued that the billion-deaths claim was a leap rather than a forecast [4]. He contended that the argument skips the causal chain between better AI tools and global catastrophe, noting that critical systems are protected by access controls, network segmentation, human approval, monitoring, and audit trails [4]. Linthicum, who advises large companies and government agencies, also asserted that artificial general intelligence remains decades away [4]. His critique addressed Gates’ earlier remark, not the new cyber forecast.

Read side by side, the two positions overlap more than the headlines suggest. Both treat today’s most pressing AI risk as a practical security problem rather than a science fiction plot [3][4].

Today’s most pressing AI risk looks like a practical security problem, not a science fiction plot.

Who Should Write the Rules for AI?

Gates has been direct about where he thinks responsibility belongs. In his CNN interview, TMZ reported, he called AI “the most dangerous technology humans ever invented” and argued that trusting AI companies to police themselves is a poor idea [1]. He compared the technology to cars, airplanes, and medications, all of which operate under regulation, and questioned why something he considers more dangerous should go without similar oversight [1].

His proposals extend into the broader economy. According to TMZ, Gates said society should seriously consider taxing AI and robots and protecting certain jobs from automation [1]. He also told Tapper he has spoken with several politicians and hopes to sit down with President Trump and his team [1]. Supporters see measures like these as a cushion for workers facing disruption. Critics would point out that new taxes on productive technology can slow investment and push innovation overseas, and that deciding which jobs deserve protection tends to favor the well-connected over the broader workforce.

Democratic strategist Oliver Libby, speaking with Fox News Digital, framed the challenge as balancing continued AI progress with smart regulation, and argued that leaders who keep the country from falling behind while making sure no American is left out will earn public trust [3]. He called for federal and international standards, platform responsibility, penalties for cyber and deepfake abuse, child safety measures, and preparation for possible workforce disruption [3].

There is common ground worth noting. Targeted penalties for criminals who weaponize AI address the specific threat Gates described without placing broad restrictions on the companies and workers building the technology. Keeping American developers competitive, a point Libby himself raised [3], also matters if the country wants to set the standards rather than follow someone else’s.

What Practical Steps Can Protect Personal Accounts Now?

Policy debates move slowly, while the threats Gates described already run on a working business model. The encouraging part is that the most effective defenses are within reach of ordinary people, and they mirror the layered approach Linthicum described for enterprise systems [4].

Access control comes first. Strong, unique passwords and multi-factor authentication on banking, brokerage, and email accounts make stolen credentials far less useful to an attacker. Email deserves special care, since it is often the key that resets every other account a person owns.

Monitoring is the second layer. Turning on transaction alerts and reviewing statements regularly shortens the window between an unauthorized charge and its discovery, which matters when intruders can move through a compromised network in under half an hour [3]. Building a routine around the best ways to track your expenses makes unfamiliar charges much easier to spot early.

Human approval is the third. Gates described extortion as a matter of threatening harm and demanding payment [3], and AI tools make those threats more convincing. Any urgent request for money, a password, or a verification code deserves a pause and an independent check through a known phone number or official website, never the contact details supplied in the message itself.

Finally, a steady long-term plan is its own form of protection. Investors who are not forced to sell during a sudden, manufactured swing give attackers far less to profit from.

The most effective defenses are within reach of ordinary people.

Final Thoughts

Bill Gates’ latest warning is narrower and more practical than some of his past statements. He is not predicting machines that slip out of human control next year. He is predicting that criminals will use powerful AI tools to run familiar schemes such as extortion and market manipulation with greater speed and polish [3]. Industry data from CrowdStrike points the same way, though it comes from a single company with a stake in the issue [3].

The bigger debates over existential risk, new taxes on AI, and how much regulation is wise will play out in Washington and beyond, and reasonable people disagree on them [1][4]. Those questions deserve careful, skeptical attention rather than headline-driven reactions.

In the meantime, the first line of defense remains personal. Strong account security, regular monitoring, and a healthy suspicion of urgent demands are within every household’s control, regardless of what policymakers decide.

Works Cited

[1] “Bill Gates Warns A.I. Is Most Dangerous Technology Ever, Wants Trump Meeting.” TMZ, 11 Oct. 2026, www.tmz.com/2026/10/11/bill-gates-warning-about-artificial-intelligence/.

[2] “Bill Gates Warns of Potential AI-Driven Cyberattacks in the Next.” CNN, 11 Oct. 2026, www.cnn.com/2026/10/11/politics/video/gates-ai-cyberattacks-sotu.

[3] Womack, CJ. “Bill Gates Sounds Alarm on AI-Powered Cyberattacks, Warns of ‘Negative Surprises’ within a Year.” Fox News, 11 Oct. 2026, www.foxnews.com/media/bill-gates-sounds-alarm-ai-powered-cyberattacks-warns-negative-surprises-within-year.

[4] Linthicum, David S. “The AI Doomsday Gap: Why ‘A Billion Deaths’ Is a Leap, Not a Forecast.” Cybersecurity Insiders, 7 Oct. 2026, www.cybersecurity-insiders.com/the-ai-doomsday-gap-why-a-billion-deaths-is-a-leap-not-a-forecast/.

[5] Womack, CJ. “Bill Gates Issues Warning on ‘Bad People’ Exploiting a Lack of AI Safeguards to Commit Cyberattacks.” New York Post, 11 Oct. 2026, nypost.com/2026/10/11/tech/bill-gates-issues-warning-on-bad-people-exploiting-a-lack-of-ai-safeguards-to-commit-cyberattacks/. Originally published in Fox News, 11 Oct. 2026.