Hackers are now turning company AI tools into weapons against the very businesses that built them.
A new global cybersecurity report released October 14, 2026, found that 84% of surveyed ransomware victims said attackers used artificial intelligence in some way during the attack, according to Semperis' 2026 Ransomware Risk Report covered by CBS News. The survey polled 1,200 IT and security professionals across nine countries and eight industries, exposing how deeply AI ransomware attacks have changed the threat landscape.
PwC Cybersecurity Budget Survey: 84% Expect Spending To Jump as AI Risks Grow
Survey Reveals Scale of AI Ransomware Attacks Worldwide
Semperis, a cybersecurity firm that focuses on protecting identity systems, built its report on responses from 1,200 professionals working in IT and security roles. The number that grabbed headlines was simple and alarming: 84% of people who had actually lived through a ransomware attack said AI played some role in it. CBS News framed this as AI being linked to 84% of ransomware attacks, though the figure reflects survey responses rather than a direct count of every attack on Earth.
Dig a little deeper into the numbers and the picture gets more specific. Forty-four percent of respondents said attackers actually exploited their own organization's AI agents or systems. Another 40% said they experienced what they described as an AI-based ransomware attack. These are three different measurements, not three versions of the same statistic, and treating them as identical would miss the real story. Attackers aren't just using AI as a tool from the outside. In many cases, they're turning a company's own AI investment into the weak point.
How Criminals Are Weaponizing Artificial Intelligence
The report lays out several ways criminals are putting AI to work. Attackers are reportedly using it to scan corporate networks faster and find weak spots that would take humans much longer to locate. They're also using AI to automate pieces of their intrusion campaigns and to craft social-engineering messages that are tailored to specific employees, making phishing emails and scam calls far more convincing.
Chris Inglis, who once served as the White House's National Cyber Director, told CBS News that AI is lowering the bar for entry into cybercrime. He said the technology lets people with limited technical skill automate attacks, hunt down vulnerabilities, and personalize deception schemes that used to require real expertise. That warning lines up with what the Semperis data shows: AI isn't just making skilled hackers faster, it's letting less skilled ones compete.
Identity Systems and Personal Threats Add New Pressure
The report also connects AI ransomware attacks to a bigger trend around identity systems. According to the findings, 87% of reported attacks compromised identity infrastructure, meaning the credentials, logins, and directories that control who can access what inside a company. When attackers get into these systems, recovery becomes much harder, because the tools companies would normally use to lock hackers out may themselves be compromised.
Beyond the technical damage, the survey points to a disturbing human cost. More than one-third of victims reported psychological or physical threats made against employees. Separately, 37% of respondents said attackers threatened employees' family members or friends, and 14% said workers or their relatives had to change jobs, offices, or even homes because of these threats. This marks a shift from attacks that only hit computers to attacks that follow people into their personal lives.
Companies Fight Back With AI, But Gaps Remain
It's not all one-sided. Around 49% of survey respondents said they used AI tools themselves to detect or defend against ransomware. That shows companies aren't sitting still while criminals adopt new technology. But the defense isn't foolproof. Twenty-three percent of respondents said their AI-based defenses actually failed to stop the malware, a reminder that having AI on your side doesn't guarantee safety.
There's also a striking repeat-attack problem. Sixty-nine percent of victims said they suffered multiple successful ransomware attacks within the past year, suggesting that many organizations are getting hit again and again rather than just once. On a slightly more hopeful note, the share of victims who paid a ransom dropped to 59%, down from 69% the year before, according to reporting on the study. That dip hints at growing resistance or better recovery options, even though paying still remains common.
What This Means for Businesses Going Forward
The Semperis findings suggest companies now have to think about AI in two very different ways: as a shield and as a possible open door. Securing AI agents, limiting what they're allowed to access, and watching identity systems closely are likely to become bigger priorities for security teams. The report doesn't point to any new law or government deadline tied to these findings, so for now the response is falling on companies themselves.
What happens next will likely involve more scrutiny of how Semperis gathered its data, along with other researchers trying to confirm whether this AI trend shows up in actual attack records, not just survey answers. Until then, the clearest takeaway is that AI ransomware attacks are no longer a future risk. For a growing share of victims, they're already here, and the lines between attacker tools and company tools are starting to blur.
