IPCRIME | THREAT INTELLIGENCE | 15 NOVEMBER 2026
Ransomware and data-extortion activity remains elevated, while smaller criminal groups are experimenting with AI-assisted deception and operational support, according to researchers interviewed for a November 14 WIRED analysis. The picture is serious, but a leak-site posting should not be counted automatically as a verified successful compromise.
What the measurements show
Security specialists cited by WIRED estimate nearly 3,000 alleged ransomware incidents on criminal leak sites during the summer, with some comparisons suggesting a 20–40 percent increase over spring. Halcyon’s analysis found roughly 111 groups claiming attacks between July and September. These are estimates based on observed claims, not a comprehensive census of confirmed incidents or victims.
Where AI fits
Analysts say AI is being used for tasks such as more convincing phishing, language localization and extortion negotiations. Their evidence does not support a simplistic claim that fully autonomous ransomware operations are now routinely running end to end without people. Experienced criminal operations still rely substantially on human decisions and existing access networks.
A fragmented criminal market
WIRED’s reporting describes new operations that make many claims in short periods, sometimes with questionable verification. Some established groups remain prominent, but researchers are also watching inexperienced operators attempt independent campaigns outside established ransomware-as-a-service structures.
What defenders can do
Protect identity systems and backups, limit administrative access, validate remote access controls, and rehearse how to respond to data-theft extortion even where files are not encrypted. An apparent decline in encryption would not necessarily mean less damage: stolen information and disrupted operations can still carry severe consequences.
Source documents and further reading
Independent IPCRIME research-based coverage. Incident dates, agency publication dates and the date of this article are distinguished above. Product guidance is not a claim of independent penetration testing.