Risky Business #826 -- A week of AI mishaps and skulduggery

On this week’s show, Patrick Gray, Adam Boileau and James WIlson discuss the week’s cybersecurity news. They cover:

• Low skill actors compromise 600 Fortinets with AI-generated playbooks

• Anthropic calls out Chinese AI firms over model distillation

• Meta’s director of AI safety tells her ClawdBot not to delete her mail… so of course it does

• Peter Williams cops 7 years in jail for selling L3 Harris Trenchant’s exploits to Russia

• Ivanti got hacked in 2021 via… bugs in Ivanti

This episode is sponsored by line-rate network capture system Corelight. CEO Brian Dye joins to discuss what AI can do for defenders, and what it can’t.

This episode is also available on Youtube (youtu.be/yy5H2eE5h98).

Show notes

• AI-augmented threat actor accesses FortiGate devices at scale (aws.amazon.com/blogs/security/ai-augmented-thr…)

• "this reads to me like: they ran existing tools.... but with a cool dashboard :D" (x.com/uk_daniel_card/status/202515819701984912…)

• Anthropic accuses Chinese labs of trying to illicitly take Claude’s capabilities | CyberScoop (cyberscoop.com/anthropic-accuses-chinese-labs-…)

• Detecting and preventing distillation attacks (anthropic.com/news/detecting-and-preventing-di…)

• Hegseth warns Anthropic to let the military use the company’s AI tech as it sees fit, AP sources say (apnews.com/article/anthropic-hegseth-ai-pentag…)

• Anthropic Rolls Out Embedded Security Scanning for Claude (cyberscoop.com/anthropic-claude-code-security-…)

• AWS's AI Coding Bot Kiro Caused a 13-Hour Outage (arstechnica.com/ai/2026/02/an-ai-coding-bot-to…)

• Running OpenClaw safely: identity, isolation, and runtime risk (microsoft.com/en-us/security/blog/2026/02/19/r…)

• Former Adobe, Cisco and Salesforce CISO talks AI pentesting (risky.biz/RBFEATURES2)

• History Repeats: Security in the AI Agent Era (risky.biz/RBFEATURES1)

• Meta Director of AI Safety Allows AI Agent to Accidentally Delete Her Inbox (404media.co/meta-director-of-ai-safety-allows-…)

• Microsoft says Office bug exposed customers' confidential emails to Copilot AI | TechCrunch (techcrunch.com/2026/02/18/microsoft-says-offic…)

• The (tangential) fix: Microsoft adds Copilot data controls to all storage locations (bleepingcomputer.com/news/microsoft/microsoft-…)

• Ex-L3Harris executive sentenced to 87 months in prison for selling zero-day exploits to Russian broker (cyberscoop.com/l3harris-executive-peter-willia…)

• Treasury Sanctions Exploit Broker Network for Theft and Sale of U.S. Government Cyber Tools (home.treasury.gov/news/press-releases/sb0404)

• Risky Bulletin: Russia starts criminal probe of Telegram founder Pavel Durov (risky.biz/risky-bulletin-russia-starts-crimina…)

• Ukraine pushes tighter Telegram regulation, citing Russian recruitment of locals (therecord.media/ukraine-telegram-regulation-ru…)

• The watchers: how openai, the US government, and persona built an identity surveillance machine that files reports on you to the feds (vmfunc.re/blog/persona)

• Persona emails customers saying they don’t work with ICE or DHS amid ‘surveillance’ claims (piunikaweb.com/2026/02/20/persona-denies-ice-d…)

• Inside the Fix: Analysis of In-the-Wild Exploit of CVE-2026-21513 (akamai.com/blog/security-research/2026/feb/ins…)

• Ivanti hacked in 2021 via its own product (bloomberg.com/news/features/2026-02-19/vpn-use…)

• Fed agencies ordered to patch Dell bug by Saturday after exploitation warning | The Record from Recorded Future News (therecord.media/fed-agencies-ordered-to-patch-…)

• From BRICKSTORM to GRIMBOLT: UNC6201 Exploiting a Dell RecoverPoint for Virtual Machines Zero-Day (cloud.google.com/blog/topics/threat-intelligen…)