Good report:

Executive Summary: Let’s say you wanted to make sure that your AI is secure. Can you just maximize the security and privacy benchmark and call it a day? Nope, because benchmarks don’t actually work for measuring AI capabilities (even when they are NOT emergent systemic properties like security). So let’s take a step back: how do you measure security in the first place? Good question. Over the last 30 years, security engineering for software evolved from black box penetration testing, through whitebox code analysis and architectural risk analysis to de facto process-driven standards like the Building Security In Maturity Model (BSIMM). Software had a very deep impact on business operations, and it appears that AI is going to have an even deeper impact. Will a software security-like measurement move work for AI? Probably. In the meantime we can make real progress in AI security by cleaning up our WHAT piles and managing risk by identifying and applying good assurance processes. (Spoiler alert: no matter what we do, we still don’t get a security meter for AI, so we need to be extra vigilant about security.)...

Not by name, but Laurie Anderson quotes me in one of the tracks of her new album:

My favorite quote is from a cryptologist who said “If you think technology will solve your problems, you don’t understand technology and you don’t understand your problems.”

Also in interviews:

“Of course, it’s ridiculous, outrageous, blah, blah, blah,” Anderson says about the ad. ‘But, I mean, my favorite quote on this is from a cryptologist who said, ‘If you think technology will solve your problems, you don’t understand technology ­ and you don’t understand your problems.’ And I think I’m completely on board with that.”...

It’s nasty, but it requires physical access to the computer:

The exploit, named YellowKey, was published earlier this week by a researcher who goes by the alias Nightmare-Eclipse. It reliably bypasses default Windows 11 deployments of BitLocker, the full-volume encryption protection Microsoft provides to make disk contents off-limits to anyone without the decryption key, which is stored in a secured piece of hardware known as a trusted platform module (TPM). BitLocker is a mandatory protection for many organizations, including those that contract with governments...

Article about the bigfin squid.

As usual, you can also use this squid post to talk about the security stories in the news that I haven’t covered.

Blog moderation policy.

Some AI-based video age-verification checks can be fooled with a fake mustache.

This is a current list of where and when I am scheduled to speak:

• I’m giving a virtual talk on “The Security of Trust in the Age of AI,” hosted by the Financial Women’s Association of New York, at 6:00 PM ET on May 21, 2026.
• I’m speaking at the Potsdam Conference on National Cybersecurity at the Hasso Plattner Institut in Potsdam, Germany. The event runs June 24–25, 2026, and my talk will be the evening of June 24.
• I’m speaking at the Digital Humanism Conference in Vienna, Austria, on Tuesday, June 26, 2026.
• I’m speaking at the ...

Last month, Anthropic made a remarkable announcement about its new model, Claude Mythos Preview: it was so good at finding security vulnerabilities in software that the company would not release it to the general public. Instead, it would only be available to a select group of companies to scan and fix their own software.

The announcement requires context—but it contained an essential truth.

While Anthropic’s model is really good at finding software vulnerabilities, so are other models. The UK’s AI Security Institute found that OpenAI’s GPT-5.5, already generally available, is comparable in capability. The company Aisle ...

The UK’s AI Security Institute evaluated GPT-5.5’s ability to find security vulnerabilities, and found that it is comparable to Claude Mythos. Note that the OpenAI model is generally available.

Here is the Institute’s evaluation of Mythos.

And here is an analysis of a smaller, cheaper model. It requires more scaffolding from the prompter, but it is also just as good.

This is the worst Linux vulnerability in years.

TL;DR

• copy.fail is a Linux kernel local privilege escalation, not a browser or clipboard attack. Disclosed by Theori on 29 April 2026 with a working PoC.
• It abuses the kernel crypto API (AF_ALG sockets) plus splice() to write four bytes at a time straight into the page cache of a file the attacker does not own.
• The exploit works unmodified across Ubuntu, RHEL, Debian, SUSE, Amazon Linux, Fedora and most others. No race condition, no per-distro offsets.
• The file on disk is never modified. AIDE, Tripwire and checksum-based monitoring see nothing. ...

Turns out that LLMs are really good at hiding text messages in other text messages.