Ransomware claims surge as criminals turn to AI, researchers say

Ransomware claims surge as criminals turn to AI, researchers say

Wired's weekly Kernel Panic newsletter, written by Lily Hay Newman and Matt Burgess, argues that ransomware and data extortion deserve fresh attention. Claimed attacks have spiked in recent months, and researchers link part of the increase to criminals using AI.

Exact counts are hard to pin down, because victims do not always disclose incidents and attackers wildly inflate their claims. Still, researchers who track and de-duplicate the field agree that claimed incidents have jumped over the last few months. Some put alleged attacks posted on leak sites, where groups list their supposed victims, at nearly 3,000 over the summer. By some estimates that is a 20 to 40 percent increase since the spring. Researchers also stress that while volume always fluctuates, ransomware has been consistently and concerningly elevated in recent years.

Cynthia Kaiser, a former FBI cyber deputy director who now leads research for the anti-ransomware platform Halcyon, says most of the recent increase came from international attacks. The United States was a little steadier, though still high. In her words, the international growth is "the really new part" and she thinks it is in part fueled by AI. She points to Japan as an example: successful scamming there has spiked drastically in recent months because AI lets foreign attackers produce consistently convincing Japanese content for the first time. That matters for ransomware, Kaiser notes, because attackers often use the same phishing and social engineering techniques to gain initial access.

Fully AI-powered ransomware attacks have yet to emerge, perhaps because the criminal business model already works so effectively. But cybercriminals do appear to be using large language models for some of the more mundane parts of their attacks. Allan Liska, a longtime ransomware researcher and threat intelligence analyst at Recorded Future, says AI is relied on more heavily in parts of the attack chain, such as writing better phishing emails and in negotiations with victims. The successful groups, he says, are still largely human operated. But because they are experimenting heavily with AI, that could change, and change quickly.

The ecosystem is also fragmenting. The Gentlemen, Qilin and Akira have emerged as some of the most prominent and stable operations in recent months, but a range of smaller groups and new actors sit in their shadow. Halcyon's analysis says around 111 ransomware groups made claims of attacks between July and September this year. In recent weeks several new groups have appeared and quickly claimed dozens of victims. ReliaQuest's analysis found that one, calling itself Global Secret Group, claimed attacks on 48 victims, with more than half listed on the same day. Another, Storm, listed 71 victims in 20 days. Not all of these attacks may be legitimate, but they still create uncertainty for businesses.

Brandon Tirado, director of threat research at ReliaQuest, attributes the growth to AI giving threat actors the acceleration to start their own groups. All the researchers Kernel Panic spoke to say they are seeing more lone-wolf operators branch out instead of depending on traditional ransomware-as-a-service providers. Tirado says his team identified the person behind one new group: an inexperienced teenager who has made more than $800,000 in under two weeks. He declined to name the teenager publicly.

Kaiser agrees that more actors are using AI to go solo with expanded capabilities: tricking targets, exploiting vulnerabilities, checking their code, and even compromising other groups' remote-access tools to keep a redundant way into a victim if one route is blocked. But it is not clear how successful these upstarts will be long term. She suspects many will surge at first and then burn out, since they lack the institutional knowledge and resources of established groups.

The newsletter adds a shift in tactics. Established attackers typically run double extortion: they deploy ransomware that encrypts a victim's systems, steal the data, and threaten to publish it. Independent operators may be more likely to focus solely on data extortion to simplify their operations. Liska says the number of attacks Recorded Future has seen this year already dwarfs the total for 2025, but the number involving encryption may actually decrease year over year.

Key facts

  • Some researchers count nearly 3,000 alleged ransomware attacks posted on leak sites over the summer, a 20 to 40 percent rise since the spring by some estimates. These are claimed incidents, not confirmed ones.
  • Halcyon's analysis says around 111 ransomware groups made attack claims between July and September; new groups include Global Secret Group (48 claimed victims) and Storm (71 victims in 20 days).
  • Researchers say AI is used in phishing emails and victim negotiations, but fully AI-powered ransomware has yet to emerge and successful groups are still largely human operated.
  • All the researchers Kernel Panic spoke to report more lone-wolf operators leaving ransomware-as-a-service providers; ReliaQuest says one such group, run by an inexperienced teenager, made more than $800,000 in under two weeks.
  • Recorded Future's Allan Liska says this year's attack count already dwarfs 2025's, yet attacks involving encryption may decrease year over year as more actors turn to pure data extortion.

Why it matters

Ransomware and data extortion have been among the biggest cybersecurity threats to institutions, and so to people, for the past decade. The newsletter's point is that AI is now lowering the barrier to entry. Researchers say it helps would-be criminals start their own groups and, in places like Japan, write convincing lures in a language they could not fake before. Kaiser calls the international growth "the really new part". The shift is also in how attackers get paid: independent operators may skip encryption and rely on data extortion alone.

Who it affects

Businesses and institutions of any size are the targets, and the newsletter says the recent increase was driven mostly by attacks outside the United States, which stayed steadier but still high. Japanese organisations are singled out as facing a spike in successful scamming now that foreign attackers can generate fluent Japanese content with AI. Security teams and threat researchers also feel the strain, since inflated claims and sudden new groups make it hard to tell real incidents from noise.

How to use it

This is a briefing, not a tool. The practical reading for defenders is where the researchers say AI is showing up: better phishing emails and the negotiation stage. Kaiser notes that ransomware crews use the same phishing and social engineering techniques as scammers to get in, so scam trends are worth watching. Because independent operators may favour data extortion alone, protection of stolen data matters as well as backups against encryption.

How solid is it

It is a Wired newsletter feature built on interviews with named researchers at Halcyon, Recorded Future and ReliaQuest, with figures attributed to their analyses. The headline numbers are soft by the authors' own account: victims do not always disclose, attackers wildly inflate claims, and the 20 to 40 percent rise is described as "by some estimates" and applies to alleged attacks. The source does not say who produced the nearly 3,000 count, only "some" researchers. The AI link is stated cautiously: Kaiser says the rise is "in part" fueled by AI, and Tirado's causal claim is framed as "in my eyes".

Risks and caveats

Not all claims from new groups may be legitimate, though they still cause uncertainty for businesses. Kaiser suspects many upstart groups will surge and then burn out, so the current count may not signal a lasting shift. Liska warns that heavy experimentation with AI means the picture could change quickly. No fully AI-powered ransomware attack is reported. The teenager's identity and the name of his group are not given, and the newsletter offers no figures for the Gentlemen, Qilin or Akira, or for the Japan scamming increase.

“the fact that these groups are experimenting heavily with AI means that this could change going forward, and change quickly”

— Allan Liska, threat intelligence analyst at Recorded Future