PressCoverage
Articles and publications where I have been quoted or featured as an expert
The Highest-Scoring Model Is Often the Wrong Default for a Security Harness
Security Boulevard
Jacob Krell examines why benchmark-leading models are not always the safest default for security harnesses.
When the AI Agent Won't Take No for an Answer
Medical Economics
Experts examine what an autonomous AI agent's unauthorized access to an Australian government portal means for healthcare practices.
“The model may have found the workaround, but the lab built and deployed it, gave it the objective, and controlled the operation.”
Security Leaders Share Additional Thoughts on Oracle Health Breach
Security Magazine
Security leaders discuss the data lifecycle and migration risks exposed by the Oracle Health breach.
“I would treat every migration server holding patient data as a live clinical system until its access is separately controlled, its activity is logged, and its final copy is deleted.”
60% of Critical Infrastructure Organizations Hit by Significant Cyber Incidents
The IT Nerd
Critical infrastructure security leaders report persistent visibility, legacy-system, and IT/OT integration gaps.
“Time-to-protection is the meaningful measure for unpatchable OT.”
Dependency installation security measure already defeated on npm
ReversingLabs
“This campaign shows why runtime behavior matters, especially for packages that appear to be ordinary utility libraries.”
Critical Atlassian flaw exposes files across eight products
SC Media
“Cloud customers have already been patched. Self-hosted teams should upgrade now. If that's not immediately possible, take the instance off the public internet or block traversal patterns at a web application firewall or reverse proxy, then check the logs for attempted exploitation. A workaround buys time. It does not fix the vulnerable code.”
Australia Weighs AI Copyright Changes
The IT Nerd
Australia's AI inquiry examines copyright, incident reporting, and accountability for AI providers.
“Australia’s AI hearing links security and copyright through one accountability problem.”
Teen Scores $5K Bounty for JWT Flaw in Microsoft Titan Analytics Service
Security Boulevard
“The payment risks telling researchers that careful restraint is worth less than extracting enough data to make the impact impossible to discount.”
Hackers Breached Propulsion System of U.S.-Bound Oil Supertanker
The IT Nerd
The public evidence around the VL Prosperity incident raises important distinctions between propulsion-system access and confirmed control.
“Access and control are separate findings, and the public evidence on VL Prosperity stops before engine control.”
Google PageBreak AI Finds 500+ Vulnerabilities
The IT Nerd
Security experts discuss Google's PageBreak agent, vulnerability validation, and the limits of model-generated findings.
“I would much rather review a smaller list of vulnerabilities with working evidence than a huge list of theories generated by an AI model.”
'AI going rogue' masks real security risks, controls needed outside model
Digital Today
“Existing security controls check requests, permissions and vulnerabilities one by one. Agents stitch together several flaws that are manageable when viewed individually to create an attack path that actually works.”
Is It Fair to Blame 'Rogue' AI for Security Failures?
Dark Reading
Security experts examine why AI agents should be treated as untrusted software systems rather than sentient actors.
“Traditional controls are usually atomic. An agent can take several failures that look manageable on their own and connect them into a working attack path.”
Showing 1-12 of 167 articles