AI-Assisted Phishing and Extortion: Ransomware Groups Accelerate Attacks Through Language Models
Cybercriminals are leveraging large language models to automate phishing, negotiation, and social engineering in ransomware campaigns, with attack volumes climbing 20–40% since spring 2026.
Last verified:
AI-Accelerated Phishing and Ransom Negotiation Drive Ransomware Surge
Cybercriminals are increasingly deploying large language models to automate components of ransomware attack chains, triggering a measurable spike in claimed incidents. According to Wired AI, ransomware groups are using AI to generate more convincing phishing emails, automate victim negotiation, and produce multilingual social engineering content—functions that previously required substantial manual effort. Researchers tracking attack volume on leak sites documented nearly 3,000 alleged ransomware attacks over summer 2026, marking a 20–40% increase since spring, though fluctuations in reporting and false claims make exact numbers difficult to verify.
The pattern reveals a critical distinction: while fully autonomous AI-powered ransomware has not yet appeared, the criminal business model is evolving incrementally through human-AI collaboration. Cynthia Kaiser, former FBI Cyber Deputy Director now leading research at anti-ransomware platform Halcyon, attributes the surge partly to international expansion enabled by AI. “What we see is most of the recent increase was fueled by international attacks,” Kaiser told Wired AI, emphasizing that language models now allow foreign threat actors to generate convincing localized content in non-English languages for the first time at scale.
Why Language Models Lower Barriers to Entry
The accessibility that AI tools provide reduces operational friction in the attack chain. Allan Liska, a ransomware researcher at Recorded Future, distinguishes between experimentation and operational maturity: “We are seeing AI relied on more heavily in certain aspects of the ransomware attack chain, such as creating better phishing emails and in the negotiation process, but the ‘successful’ ransomware groups are still largely human operated.” The implication is that while leading criminal organizations have not yet fully delegated decision-making to AI, their active experimentation signals a readiness to deepen automation if the approach proves effective.
The connection between broader scamming trends and ransomware is direct. Threat actors use identical social engineering and phishing techniques to gain initial network access; in Japan, for instance, AI-generated content in Japanese has enabled foreign scammers to operate at scale in a market previously defended by language barriers. The same advantage applies to ransomware campaigns targeting international victims.
Why This Matters
The data suggests that ransomware remains a structurally attractive criminal enterprise precisely because the traditional human-operated model already yields substantial returns. AI is not driving adoption of ransomware itself—it is lowering the cost and complexity of specific subtasks, making the existing playbook more efficient and scalable. For security teams, the immediate implication is that phishing emails will become harder to distinguish from legitimate communication, and negotiation messages will exhibit fewer linguistic tells that reveal foreign threat actors. The long-term risk, flagged by Liska and Kaiser, is that if fully autonomous orchestration proves viable, the speed and volume of attacks could exceed current detection and response capabilities. Organizations should assume that ransomware groups—particularly Gentlemen, Qilin, and Akira, which are documented experimenting with AI—will gradually shift from testing to production deployment as confidence in automated tactics increases.
Frequently Asked Questions
Are ransomware attacks fully automated by AI yet?
No. Fully AI-powered ransomware campaigns have not yet emerged; leading criminal groups remain largely human-operated. However, they are experimenting heavily with language models for specific tasks like phishing and ransom negotiation, which researchers say could change the threat landscape rapidly.
How much has ransomware volume increased?
Claimed incidents on leak sites rose to nearly 3,000 over summer 2026, representing a 20–40% increase since spring, according to researchers tracking the field. The uptick is concentrated internationally rather than in the United States.
Why are ransomware groups experimenting with AI?
AI enables faster generation of convincing phishing emails in multiple languages and reduces manual effort in victim negotiation—tasks that previously required human judgment. The technology also improves social engineering by producing culturally and linguistically plausible content at scale.
Which ransomware groups are known to use AI tactics?
Gentlemen, Qilin, and Akira are among groups documented using AI-assisted techniques, particularly for phishing and extortion messaging.