This seems like an important advance in LLM security against prompt injection:

Google DeepMind has unveiled CaMeL (CApabilities for MachinE Learning), a new approach to stopping prompt-injection attacks that abandons the failed strategy of having AI models police themselves. Instead, CaMeL treats language models as fundamentally untrusted components within a secure software framework, creating clear boundaries between user commands and potentially malicious content.

[…]

To understand CaMeL, you need to understand that prompt injections happen when AI systems can’t distinguish between legitimate user commands and malicious instructions hidden in content they’re processing...

The company doesn’t keep logs, so couldn’t turn over data:

Windscribe, a globally used privacy-first VPN service, announced today that its founder, Yegor Sak, has been fully acquitted by a court in Athens, Greece, following a two-year legal battle in which Sak was personally charged in connection with an alleged internet offence by an unknown user of the service.

The case centred around a Windscribe-owned server in Finland that was allegedly used to breach a system in Greece. Greek authorities, in cooperation with INTERPOL, traced the IP address to Windscribe’s infrastructure and, unlike standard international procedures, proceeded to initiate criminal proceedings against Sak himself, rather than pursuing information through standard corporate channels...

Text “SQUID” to 1-833-SCI-TEXT for daily squid facts. The website has merch.

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

Long story of a $250 million cryptocurrency theft that, in a complicated chain events, resulted in a pretty brutal kidnapping.

Interesting:

The company has released a working rootkit called “Curing” that uses io_uring, a feature built into the Linux kernel, to stealthily perform malicious activities without being caught by many of the detection solutions currently on the market.

At the heart of the issue is the heavy reliance on monitoring system calls, which has become the go-to method for many cybersecurity vendors. The problem? Attackers can completely sidestep these monitored calls by leaning on io_uring instead. This clever method could let bad actors quietly make network connections or tamper with files without triggering the usual alarms...

Interesting research: “Guillotine: Hypervisors for Isolating Malicious AIs.”

Abstract:As AI models become more embedded in critical sectors like finance, healthcare, and the military, their inscrutable behavior poses ever-greater risks to society. To mitigate this risk, we propose Guillotine, a hypervisor architecture for sandboxing powerful AI models—models that, by accident or malice, can generate existential threats to humanity. Although Guillotine borrows some well-known virtualization techniques, Guillotine must also introduce fundamentally new isolation mechanisms to handle the unique threat model posed by existential-risk AIs. For example, a rogue AI may try to introspect upon hypervisor software or the underlying hardware substrate to enable later subversion of that control plane; thus, a Guillotine hypervisor requires careful co-design of the hypervisor software and the CPUs, RAM, NIC, and storage devices that support the hypervisor software, to thwart side channel leakage and more generally eliminate mechanisms for AI to exploit reflection-based vulnerabilities. Beyond such isolation at the software, network, and microarchitectural layers, a Guillotine hypervisor must also provide physical fail-safes more commonly associated with nuclear power plants, avionic platforms, and other types of mission critical systems. Physical fail-safes, e.g., involving electromechanical disconnection of network cables, or the flooding of a datacenter which holds a rogue AI, provide defense in depth if software, network, and microarchitectural isolation is compromised and a rogue AI must be temporarily shut down or permanently destroyed. ...

Android phones will soon reboot themselves after sitting idle for three days. iPhones have had this feature for a while; it’s nice to see Google add it to their phones.

A live colossal squid was filmed for the first time in the ocean. It’s only a juvenile: a foot long.

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

Discord is testing the feature:

“We’re currently running tests in select regions to age-gate access to certain spaces or user settings,” a spokesperson for Discord said in a statement. “The information shared to power the age verification method is only used for the one-time age verification process and is not stored by Discord or our vendor. For Face Scan, the solution our vendor uses operates on-device, which means there is no collection of any biometric information when you scan your face. For ID verification, the scan of your ID is deleted upon verification.”...

Mitre’s CVE’s program—which provides common naming and other informational resources about cybersecurity vulnerabilities—was about to be cancelled, as the US Department of Homeland Security failed to renew the contact. It was funded for eleven more months at the last minute.

This is a big deal. The CVE program is one of those pieces of common infrastructure that everyone benefits from. Losing it will bring us back to a world where there’s no single way to talk about vulnerabilities. It’s kind of crazy to think that the US government might damage its own security in this way—but I suppose no crazier than any of the other ways the US is working against its own interests right now...

As AI coding assistants invent nonexistent software libraries to download and use, enterprising attackers create and upload libraries with those names—laced with malware, of course.

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

• I’m giving an online talk on AI and trust for the Weizenbaum Institute on April 24, 2025 at 2:00 PM CEST (8:00 AM ET).

The list is maintained on this page.

The Wall Street Journal has the story:

Chinese officials acknowledged in a secret December meeting that Beijing was behind a widespread series of alarming cyberattacks on U.S. infrastructure, according to people familiar with the matter, underscoring how hostilities between the two superpowers are continuing to escalate.

The Chinese delegation linked years of intrusions into computer networks at U.S. ports, water utilities, airports and other targets, to increasing U.S. policy support for Taiwan, the people, who declined to be named, said.

The admission wasn’t explicit:...

Researchers are trying to use squid color-changing biochemistry for solar tech.

This appears to be new and related research to a 2019 squid post.

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

Microsoft is reporting that its AI systems are able to find new vulnerabilities in source code:

Microsoft discovered eleven vulnerabilities in GRUB2, including integer and buffer overflows in filesystem parsers, command flaws, and a side-channel in cryptographic comparison.

Additionally, 9 buffer overflows in parsing SquashFS, EXT4, CramFS, JFFS2, and symlinks were discovered in U-Boot and Barebox, which require physical access to exploit.

The newly discovered flaws impact devices relying on UEFI Secure Boot, and if the right conditions are met, attackers can bypass security protections to execute arbitrary code on the device...

Imagine that all of us—all of society—have landed on some alien planet and need to form a government: clean slate. We do not have any legacy systems from the United States or any other country. We do not have any special or unique interests to perturb our thinking. How would we govern ourselves? It is unlikely that we would use the systems we have today. Modern representative democracy was the best form of government that eighteenth-century technology could invent. The twenty-first century is very different: scientifically, technically, and philosophically. For example, eighteenth-century democracy was designed under the assumption that travel and communications were both hard...

Neiman Lab has some good advice on how to leak a story to a journalist.

At a Congressional hearing earlier this week, Matt Blaze made the point that CALEA, the 1994 law that forces telecoms to make phone calls wiretappable, is outdated in today’s threat environment and should be rethought:

In other words, while the legally-mandated CALEA capability requirements have changed little over the last three decades, the infrastructure that must implement and protect it has changed radically. This has greatly expanded the “attack surface” that must be defended to prevent unauthorized wiretaps, especially at scale. The job of the illegal eavesdropper has gotten significantly easier, with many more options and opportunities for them to exploit. Compromising our telecommunications infrastructure is now little different from performing any other kind of computer intrusion or data breach, a well-known and endemic cybersecurity problem. To put it bluntly, something like Salt Typhoon was inevitable, and will likely happen again unless significant changes are made...

In “Secrets and Lies” (2000), I wrote:

It is poor civic hygiene to install technologies that could someday facilitate a police state.

It’s something a bunch of us were saying at the time, in reference to the vast NSA’s surveillance capabilities.

I have been thinking of that quote a lot as I read news stories of President Trump firing the Director of the National Security Agency. General Timothy Haugh.

A couple of weeks ago, I wrote:

We don’t know what pressure the Trump administration is using to make intelligence services fall into line, but it isn’t crazy to ...

The Brooklyn indie art-punk group, Two-Man Giant Squid, just released a new album.

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