Topic
Criminal use of AI tools and models to build, obfuscate, and scale malware, phishing kits, and victim-targeting across cybercrime operations.
Microsoft counted more than a million of them between August 3 and 5, each carrying a fabricated ServiceNow invoice and a staged executive email thread, with the attacker's beneficiary bank details changing between targets.
Reality
RatHat's malware console now feeds stolen texts to Google's Gemini to estimate each victim's bank balance and rank who to rob first, security firm Cleafy says. Cleafy has traced nearly 100 deployments since April 2026 and found nothing that moves money.
Reality
Malwarebytes says the fake Avast renewal page it found aimed at Belgian users was noticeably more polished than most scam sites, with clean French copy, an Active status badge and a form that asks only for a name, an email address and a phone number.
Reality
Microsoft says the phishing-as-a-service platform reached more than 12,000 inboxes at over 10,000 organizations in seven months by abusing a legitimate OAuth flow, and its Digital Crimes Unit has now disrupted the infrastructure behind the service.
Reality
Seemant Sehgal, whose firm sells continuous penetration testing, says AI agents have cut the skill needed to run a full intrusion campaign. His evidence is two August 2025 incidents and one unmeasured claim about speed.
Reality
Rapid7 recovered an exposed server running an active vishing operation. Alongside the phishing kit sat the operator's development trail, including a custom jailbreak written after one model pushed back.
Reality