SecurityWeekEduard Kovacs
Microsoft Disrupts AI Phishing Platform After 12,000 Accounts Compromised
Part of the EvilTokens story · since 23 Sep
- Wed 23 Sep, 11:38 UTC. SecurityWeek, first report. Microsoft Disrupts AI Phishing Platform After 12,000 Accounts Compromised. This article, 1st of 1.
Microsoft says it has disrupted EvilTokens, an AI-powered phishing platform that emerged in February 2026 and was used to compromise more than 12,000 email accounts across over 10,000 organisations. The affected organisations were located in the US, Canada, the UK, Australia, India and France. EvilTokens supported device-code phishing, in which attackers persuaded victims to enter a code supplied through a lure into a browser.
Completing the process authenticated the attacker’s device and could provide access tokens allowing persistent access to the victim’s email without revealing a password.
According to Microsoft, the service used AI to create targeted phishing emails, select from 44 themes for malicious messages and phishing pages, and analyse compromised inboxes to identify valuable information, potential targets and people to impersonate. Microsoft also believes AI was used to build the platform itself. Criminals reportedly paid $1,500 for initial access and $500 per month to use it.
Microsoft said it seized 50 websites used to operate EvilTokens and disabled more than 150 other domains associated with its infrastructure. It also reported that two men suspected of involvement, Felix Utomi and Waidi Segun Adams, had been arrested in the UK. The suspects were named in a Microsoft complaint that also targets five unnamed individuals; the allegations have not been presented as convictions. SpyCloud, TRM Labs, Coinbase, Health-ISAC, Cloudflare, OpenAI, Railway and the Shadowserver Foundation contributed to the takedown.
How the story developed
1 article from 1 outlet since 23 Sep. Quiet since 23 Sep. This was the first report.
- Wed 23 Sep, 11:38 UTC. SecurityWeek, first report. Microsoft Disrupts AI Phishing Platform After 12,000 Accounts Compromised. This article, 1st of 1.
- No CVE
- Larger: an outlet’s first report
- Coverage over time
- Quiet time, folded
Show all 1 articles as a listHide the list
- Wed 23 SepDay 01
- Microsoft Disrupts AI Phishing Platform After 12,000 Accounts Compromised11:38 · SecurityWeek · First report · This article, 1st of 1