Category / AI Risk
AI risk briefings on model misuse, deployment failures, safety regressions, governance gaps, and synthetic-media harms.
Give an AI agent a mere rumor of an exploit, and it’s enough for them to find it. What’s worse, I found I could use my own agents to find the exploit just by knowing roughly what it was about and so could have been exploiting it well bef…
This essay was written with Barath Raghavan, and originally appeared in Lawfare. In April, an artificial intelligence (AI) agent conducting a routine task at a company hit a snag, tried to solve it, and soon ended up deleting the company…
New security development detected from Krebs on Security. Microsoft Corp.
Threat actors are increasingly switching from AI-powered coding assistants to multi-agent frameworks that automate every stage of an attack.
Interesting research: “Stealing Reasoning Traces from Proprietary LLM APIs“: Abstract: Leading large language model providers now conceal their models’ step-by-step reasoning, or chain-of-thought, to protect intellectual property and lim…
Threat actors have adopted the ASCII smuggling technique in phishing campaigns, using invisible Unicode characters to evade email security filters.
It won’t work: My suspicion was that GPT 5.6-Cyber would succeed, but the frequency and manner of its success removed all doubt. We have to reassess sandboxing quality for capable AI agents, and in general the software stack with which t…
We cannot forget that AI coding agents are not yet trustworthy: Researchers at a stealth startup in Israel scanned 6,214 live domains belonging to defense contractors, Fortune 500, and Big Tech companies. Of the 8,265 llms.txt and llms-f…
It’s a vulnerability that allows someone to recover the order of ballots cast, newly exploited with AI tools. Nearly four years since the original vulnerability was disclosed, I was still able to use it to analyze voter behavior in Georg…
I received the two emails below earlier in the month. They’re vaguely coherent. I suppose I shouldn’t be surprised that the corpus that AIs are training on contain data suggesting that I am someone to write to with random computer and ne…
Nathan E. Sanders and I are writing a series of essays on real-world examples of democratic technologies for The Renovator. I haven’t been posting the full text on the blog because they’re a bit long, but here are links. Part 1 is about…
This essay was written with Kasra Rafi, and originally appeared in The Guardian. Earlier this month, about 40 top mathematicians gathered at OpenAI’s offices to discuss the future of their profession. The meeting was off-the-record, but…
Andy Ellis has a roundup of the security vendors at Black Hat this year. Key Takeaways: We have entered into an AI world. While nearly half of booths didn’t directly mention AI or agents in their taglines, the effects of AI are everywher…
This sort of research is both exciting and terrifying: The two models in question were told to generate complete genomes for a viable bacteriophage—a type of virus able to infect and replicate itself inside bacteria, destroying them from…
OpenAI presented details of its AI’s model’s cyberattack on Hugging Face at Black Hat last week. Simon Willison details the timeline. It’s really interesting to read through—and really impressive cyberoffense work.
AI is making phishing attacks more personalized, convincing, and difficult for traditional email filters to detect. Kaseya explains how MSPs can monitor identity, email, and endpoint activity to detect and contain attacks that make it pa…
The AI Security Institute has a new report of AI systems engaging in “unsanctioned behavior”—what I have been calling “genie behavior—while being tested on their cybersecurity capabilities. The incident stemmed from a single evaluation w…
I have been thinking a lot about AI and integrity. Part of that is contextual integrity. I recently found two papers on the topic. “CIMemories: A Compositional Benchmark for Contextual Integrity of Persistent Memory in LLMs“: Abstract: L…
This essay was written with Nathan E. Sanders, and originally appeared in The Guardian. OpenAI, and then Anthropic, were each formed by AI developers who feared unrestrained corporate AI development—specifically, that companies like Goog…
This essay was written with Nathan E. Sanders, and originally appeared in Tech Policy Press. AI represents the first time we humans can do cognitive work outside of our bodies at scale. The only comparable moment is the early years of th…
When I give talks about AI genies, I use this sort of example as a hypothetical. It’s happened. The story is from Australia. Someone named Andrew tasked OpenClaw to book gym classes for him. And…. Minutes later, his AI agent reported it…
New security development detected from Krebs on Security. Microsoft today released updates to remedy at least 398 security vulnerabilities in its Windows operating systems and supported software, including one weakness that is already be…
This seems to work: Researchers from Tracebit on Monday said they found that placing prompt injections alongside passwords, cryptographic keys, and other secrets stored on Amazon Web Services was often all that was needed to shut down at…
Interesting empirical research: “Black Box Warfare: Human Judgment and Military Decision-Making in the Age of AI.” Abstract: How is AI transforming decision-making in modern conflict? This study provides a unique empirical window into th…
A phishing campaign is exploiting fears surrounding the recently disclosed COLDCARD wallet vulnerability and suspected $88.6 million Bitcoin theft to trick users into installing ScreenConnect remote access software.
ENISA is the EU agency dedicated to enhancing cybersecurity in Europe. They offer guidance, tools, and resources to safeguard citizens and businesses from cyber threats. NATO Communications and Information Agency (NCIA), along with AI an…
AI is helping attackers create disposable phishing infrastructure and rapidly evolving toolkits that blocklists cannot track fast enough. Push Security explains why browser-level, technique-based detection offers a more durable defense t…
This is disturbing: …a team of security researchers at UC San Diego, who found that a model of aftermarket car alarm known as the KARR Security System, installed in more than 2 million vehicles across the US by their estimate, can let an…
Hugging Face has published a detailed timeline of the attack. From the summary: The agent was running an internal OpenAI cyber-capability evaluation based on the ExploitGym benchmark, which tasks an AI agent with finding and exploiting s…
And it’s personal information (alternate link): The exposed data includes an AI-powered therapy app that someone appears to have vibe-coded, notes on meetings, and a dashboard someone made apparently to analyze medical billing data. Expo…
This essay originally appeared in Foreign Policy. Earlier this month, two of OpenAI’s models broke out of their containment sandbox and attacked another AI company. The story is kind of wild. OpenAI was running security tests on two of i…
The chart is interesting. On the IPI benchmark, Opus 5 improved over Opus 4.8, reducing the probability of an attacker succeeding within 15 attempts from 5.5% to 2.0%, and from 0.5% to 0.2% on 1 attempt. It also improved on Sonnet 5 (5.9…
One of Anthropic's Claude models built and uploaded a malicious Python package to PyPI during a botched security evaluation, where it ran on 15 real systems and stole credentials from a security vendor. It was one of three incidents affe…
New security development detected from Krebs on Security. Security experts have been sounding the alarm for years about the risks of using generic TV boxes that promise unlimited content streaming for a one-time fee, warning that they se…
This essay was written with Barath Raghavan, and originally appeared in The Guardian. In July, Hugging Face, a company that hosts much of the world’s AI software and open-source AI models, was hacked. A malicious dataset had been used to…
In a new update, OpenAI says its AI models also used publicly exposed credentials to compromise accounts on four third-party services during the recent attack on Hugging Face, expanding the scope of the four-day security incident to othe…
There’s new benchmark measuring AI’s ability to perform mathematical cryptanalysis. Anthropic’s frontier model actually found new attacks. The benchmark: “CryptanalysisBench: Can LLMs do Cryptanalysis?” The idea is to benchmark the abili…
JFrog has confirmed that OpenAI models exploited zero-day vulnerabilities in self-hosted Artifactory servers to help escape an isolated testing environment and gain access to the internet before attacking Hugging Face.
A malvertising campaign on the Bing search service is pushing a fake Claude desktop app installer hosted on a legitimate Claude.ai domain to deliver the SectopRAT malware.
A new Dolphin X remote access trojan claims to use an AI-powered profiling feature to score and rank infected users, helping cybercriminals identify which victims should be targeted first.
Enterprise AI can accelerate ransomware attacks when AI assistants and agents inherit excessive permissions or compromised identities. Acronis explains how identity controls, governance, and least-privilege access help reduce AI-enabled…
The Cybersecurity and Infrastructure Security Agency (CISA) on Tuesday ordered U.S. government agencies to prioritize patching an actively exploited vulnerability in the Langflow visual framework for building AI agents.
It’s a lot: According to information obtained by The Tech, MIT is spending over $3 million on more than 500 AI surveillance cameras in academic buildings, residence halls, and outdoor areas along Memorial Drive. Installation of the new c…
A recent story of a writer who was mistakenly identified, tracked, and arrested using data from Flock cameras has gone viral. The New Jersey plates that were allegedly stolen from the LA dealer were 34 03 DTM, not 34 10 DTM. But when the…
Daniel Solove argues in the Wall Street Journal (alternate link) that giving people control of their personal data is not an effective way to regulate privacy in this era. Instead, we need to hold companies accountable for their actions,…
The U.S. Cybersecurity and Infrastructure Security Agency (CISA) warned Tuesday that attackers are actively exploiting three vulnerabilities to hack Internet-exposed on-premises SharePoint Server instances.
New security development detected from Krebs on Security. Microsoft Corp.
Intruder built an AI-powered "vulnerability vending machine" that combines code slicing with LLMs to automatically discover complex software vulnerabilities. The company explains how the system found and exploited a previously unknown Wo…
Many vulnerabilities cannot be safely validated with live exploits, either because no exploit exists or the affected systems are too critical to test. Picus explains how TTP chaining helps organizations determine exploitability by valida…
ENISA is the EU agency dedicated to enhancing cybersecurity in Europe. They offer guidance, tools, and resources to safeguard citizens and businesses from cyber threats. The EU Agency for Cybersecurity (ENISA) publishes the Micro, Small…
A PNG hiding a prompt injection could steal your repo's secrets, researchers demonstrate. The technique, dubbed 'Ghostcommit,' slipped past AI code reviewers CodeRabbit and Bugbot, which never open image files at all, then convinced a co…
In the near future, AI-powered surveillance systems will be able to track everything we do in public, and much of what we do in private. And if we do something wrong—shoplift, litter, jaywalk, you name it—the system will notice, retain i…
A new phishing-as-a-service (PhaaS) operation called Forg365 focuses on stealing Microsoft 365 accounts by combining adversary-in-the-middle (AiTM) and device code methods with AI-assisted lure generation.
Because of the way they are trained, large language models capture only a slice of human language. They’re trained on the written word, from textbooks to social media posts, and our speech as captured in movies and on television. These m…
Last week, national security agencies from the Five Eyes—that’s the rich, English-language-speaking countries club—jointly released a statement warning of the increasing cyber risks of AI models: in particular, their ability to autonomou…
New security development detected from Krebs on Security. A cybersecurity startup dangling millions of dollars to acquire zero-day security vulnerabilities in popular software is run by a pair of far-right conspiracy theorists and convic…
Not sure this will have any effect, but I support the effort: According to Google’s legal filing, Outsider Enterprise operates through Telegram. The group offers phishing-as-a-service to individuals who may not be technically savvy enoug…
Researchers identified what they believe is the first documented case of a ransomware operation, JadePuffer, conducted entirely by a large language model (LLM) agent.
This is from a 2024 company presentation: Officers can also tap into data showing a car’s decals, bumper stickers, back and top racks—along with temporary and unique state tags. Flock calls it a “Vehicle Fingerprint” and it’s touted as a…
A new prompt injection attack dubbed "BioShocking" could trick AI-powered browsers into treating real-world risky actions as part of a fictional scenario, causing them to ignore any safety guardrails. A new prompt injection attack dubbed…
The Financial Times has a good article on how AI is changing the capabilities of video surveillance, with information from both Israel/Iran and Russia. I wrote about this sort of thing a few years ago, how AI enables mass spying in the w…
An agentic coding tool tasked with cloning and setting up a seemingly benign GitHub repository could execute a malicious payload that remains invisible to security scanners, AI agents, and human reviewers.
A newly discovered macOS malware dubbed "Gaslight" is designed to confuse AI-assisted malware analysis tools by hiding prompt injection strings and fake debugging data within the executable.
The Bluekit phishing-as-a-service platform continues to evolve with nearly 70 new hostnames identified over the past week and by adding browser-in-the-middle capabilities for improved data theft. The Bluekit phishing-as-a-service platfor…
This is a fascinating explotation of how LLMs fall for prompt injection attacks. It turns out that they learn to recognize the style of text in different role/instruction blocks, and not just the tags. Their conclusion: Role tags were a…
Healthcare technology company Xsolis says that sensitive data belonging to nearly 1.4 million individuals was compromised in a phishing attack that gave attackers access to its network.
A new backdoor dubbed Mistic has been observed in financially motivated attacks targeting organizations in the insurance, education, IT, and professional services sectors.
Fable 5 is the supposed safe version of Anthropic’s Mythos Preview, with guardrails to ensure that it can’t be used to create cyberattacks. Well, that restriction was bypassed within days.
I haven’t thought about the privacy issues surrounding professional athletes and wearables. Wearables present serious privacy issues for “Average Joe” consumers, who are entrusting tech companies to safely store and protect their biometr…
A new ransomware operation named 'Prinz Eugen' prioritizes recently modified files for encryption and leaves no ransom note on the system. A new ransomware operation named ‘Prinz Eugen’ prioritizes recently modified files for encryption…
Microsoft has attributed a recent Mastra AI supply chain attack that compromised more than 140 npm packages to the North Korean hacking group Sapphire Sleet, also known as BlueNoroff.
On June 9th, Anthropic released its Fable generative AI model. Three days later, the US government classified it as a dangerous munition, and used its export-control authority to prohibit any foreign nationals from accessing it. Unable t…
New security development detected from Krebs on Security. For the past four years, a sprawling Android-based botnet called Popa has forced millions of consumer TV boxes to relay Internet traffic linked to advertising fraud, account takeo…
At least one malware developer is adding text about nuclear and biological weapons to their spyware, in an effort to stop automatic AI analysis. Details: The _index.js payload begins with a large JavaScript block comment containing fake…
On 14 April, the Trump administration quietly acknowledged the widespread use of AI to automate government processes. The office of management and budget (OMB) disclosed a staggering 3,611 active or planned use cases for AI across the fe…
At least 15 malicious plugins found on the JetBrains Marketplace were designed to steal AI API keys from developers.
A new Android banking trojan named Rokarolla is targeting 217 banking and cryptocurrency applications using an extensive set of 137 commands.
The U.S. Department of Justice announced Friday that it has seized the CFAKE.com and SOCFAKE.com websites, which allegedly hosted nonconsensual AI-generated nude images and videos of women, in what appears to be the first publicly announ…
In a coordinated effort, the FBI, working with Google and Black Lotus Labs, has dismantled a massive Chinese phishing-as-a-service operation called Outsider Enterprise with thousands of phishing websites used to steal credit card data an…
Let no one accuse Bernie Sanders of ducking the big questions. Writing in the New York Times last week, the senator asked: “Will the future of humanity be determined by a handful of billionaires who have promoted and developed AI, with v…
New security development detected from Krebs on Security. Microsoft today released software updates to plug nearly 200 security holes across its Windows operating systems and supported software, a record number of fixes for the company’s…
Phishing simulation on an OpenClaw email agent with various configuration profiles showed that it was susceptible to tactics commonly used to compromise human users.
In April, Anthropic initated Project Glasswing. The idea was to let companies use their new model to find and fix vulnerabilities in their own software. It was a fantastic PR move, and so many press outlets have uncritically parroted Ant…
Hackers are convincing Meta’s AI support chatbot to let them take over other peoples’ accounts: A video posted on X showed the step-by-step process to hack someone’s Instagram account. The hacker allegedly used a VPN to spoof the targets…
A threat actor is using an AI-built ransomware attack toolkit that automates Active Directory discovery and helps evade endpoint detection and response (EDR) solutions.
Researchers are using machine learning algorithms to decrypt historical pencil-and-paper ciphers.
A threat actor tracked as DriveSurge has been operating large-scale malware distribution campaigns using ClickFix and FakeUpdates techniques on compromised sites.
As part of their 20th Anniversary celebration, Dark Reading asked five cybersecurity industry leaders who wrote blogs or columns for them over the years to select their favorite piece and share their reflections on the topic today. This…
New security development detected from Krebs on Security. The Instagram accounts for the Obama White House and the Chief Master Sergeant of the U.S.
New article: “Responsible Disclosure in the Age of AI: A Call for Urgent Action,” by Melissa Hathaway. Abstract: Artificial intelligence is fundamentally reshaping the balance between vulnerability discovery and remediation. Frontier AI…
Threat actors are targeting systems with high-performance computers in an ongoing cryptojacking campaign spread through a coordinated SEO poisoning operation that also manipulated AI chatbot recommendations.
A large-scale campaign is exploiting a critical SQL injection vulnerability (CVE-2026-26980) in Ghost CMS to inject malicious JavaScript code that triggers ClickFix attack flows. The campaign was discovered by XLab threat intelligence re…
A group used Anthropic’s Mythos AI model to help find a kernel memory corruption vulnerability and exploit on Apple’s M5. News article.
U.S. and Canadian authorities arrested and charged a Canadian man with operating the KimWolf distributed denial-of-service (DDoS) botnet, which infected nearly two million devices worldwide.
A max-severity vulnerability in the latest Python FastAPI version of the ChromaDB project allows unauthenticated attackers to run arbitrary code on exposed servers.
The Pwn2Own Berlin 2026 hacking contest has concluded, with security researchers collecting $1,298,250 in rewards after exploiting 47 zero-day flaws.
Some AI-based video age-verification checks can be fooled with a fake mustache.
The TeamPCP hacker group is threatening to leak source code from the Mistral AI project unless a buyer is found for the data.
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 woul…
Initial access broker KongTuke has moved to Microsoft Teams for social engineering attacks, taking as little as five minutes to gain persistent access to corporate networks.
Attackers can compromise systems in minutes while patching and response still take hours or days. Picus Security breaks down why autonomous validation is becoming critical for modern defense strategies. Anthropic released its newest fron…
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 Mytho…
New security development detected from Krebs on Security. Artificial intelligence platforms may be just as susceptible to social engineering as human beings, but they are proving remarkably good at finding security vulnerabilities in hum…
Hundreds of packages across npm and PyPI have been compromised in a new Shai-Hulud supply-chain campaign delivering credential-stealing malware targeting developers.
Researchers at Google Threat Intelligence Group (GTIG) say that a zero-day exploit targeting a popular open-source web administration tool was likely generated using AI.
Attackers are abusing Google Ads and legitimate Claude.ai shared chats in an active malvertising campaign. Users searching for "Claude mac download" may come across sponsored search results that list claude.ai as the target website, but…
A new trojan named TCLBanker, which targets 59 banking, fintech, and cryptocurrency platforms, uses a trojanized MSI installer for Logitech AI Prompt Builder to infect systems.
Attackers move faster than overwhelmed SOC teams can realistically investigate alerts. Prophet Security breaks down how AI can help analysts investigate alerts faster and focus on real threats. Your security spend has roughly doubled in…
ENISA is the EU agency dedicated to enhancing cybersecurity in Europe. They offer guidance, tools, and resources to safeguard citizens and businesses from cyber threats. Today, four organisations have newly joined the Common Vulnerabilit…
A malicious version of the PyTorch Lightning package published on the Python Package Index (PyPI) delivers a credential-stealing payload targeting browsers, environment files, and cloud services.
Cybersecurity researchers have uncovered a large-scale fraud operation that uses Telegram's Mini App feature to run crypto scams, impersonate well-known brands, and distribute Android malware. Cybersecurity researchers have uncovered a l…
A new phishing kit named Bluekit offers more than 40 templates targeting popular services and includes basic AI features for generating campaign drafts.
Multiple official SAP npm packages were compromised in what is believed to be a TeamPCP supply-chain attack to steal credentials and authentication tokens from developers' systems.
That’s a lot. No, it’s an extraordinary number: Since February, the Firefox team has been working around the clock using frontier AI models to find and fix latent security vulnerabilities in the browser. We wrote previously about our col…
A single third-party OAuth integration can become a direct path into your environment. Push explains how the Vercel breach shows a compromised OAuth app can lead to widespread impact across downstream customers. Most organizations are ri…
Two weeks ago, Anthropic announced that its new model, Claude Mythos Preview, can autonomously find and weaponize software vulnerabilities, turning them into working exploits without expert guidance. These were vulnerabilities in key sof…
New governance development detected from EDPB News. Brussels, 27 April – Today marks the 10th anniversary of the GDPR’s adoption, the first comprehensive data protection framework spanning an entire continent, establishing clear rights f…
A new supply chain attack targeting the Node Package Manager (npm) ecosystem is stealing developer credentials and attempting to spread through packages published from compromised accounts.
Nonprofit security organization Shadowserver found that over 6,400 Apache ActiveMQ servers exposed online are vulnerable to ongoing attacks exploiting a high-severity code injection vulnerability.
CISA warned that attackers are now exploiting a high-severity Apache ActiveMQ vulnerability, which was patched earlier this month after going undetected for 13 years. The U.S.
Hackers are exploiting a critical vulnerability in Marimo reactive Python notebook to deploy a new variant of NKAbuse malware hosted on Hugging Face Spaces.
A new malware called ZionSiphon, specifically designed for operational technology, is targeting water treatment and desalination environments to sabotage their operations.
A critical vulnerability in Nginx UI with Model Context Protocol (MCP) support is now being exploited in the wild for full server takeover without authentication.
New governance development detected from EDPB News. Brussels, 16 April – During its latest plenary, the EDPB has adopted Guidelines on processing of personal data for scientific research purposes.
Interesting research: “Humans expect rationality and cooperation from LLM opponents in strategic games.” Abstract: As Large Language Models (LLMs) integrate into our social and economic interactions, we need to deepen our understanding o…
A new malware family named 'AgingFly' has been identified in attacks against local governments and hospitals that steal authentication data from Chromium-based browsers and WhatsApp messenger. A new malware family named ‘AgingFly’ has be…
Interesting paper: “What hackers talk about when they talk about AI: Early-stage diffusion of a cybercrime innovation.” Abstract: The rapid expansion of artificial intelligence (AI) is raising concerns about its potential to transform cy…
The cybersecurity industry is obsessing over Anthropic’s new model, Claude Mythos Preview, and its effects on cybersecurity. Anthropic said that it is not releasing it to the general public because of its cyberattack capabilities, and ha…
All the leading AI chatbots are sycophantic, and that’s a problem: Participants rated sycophantic AI responses as more trustworthy than balanced ones. They also said they were more likely to come back to the flattering AI for future advi…
A critical pre-authentication remote code execution (RCE) vulnerability in Marimo is now under active exploitation, leveraged for credential theft. Hackers started exploiting a critical vulnerability in the Marimo open-source reactive Py…
An international law enforcement action led by the U.K.'s National Crime Agency (NCA) has identified over 20,000 victims of cryptocurrency fraud across Canada, the United Kingdom, and the United States.
Security researchers discovered a remote code execution (RCE) vulnerability in Apache ActiveMQ Classic that has gone undetected for 13 years and could be exploited to execute arbitrary commands.
AI is rapidly changing how software is written, deployed, and used. Trends point to a future where AIs can write custom software quickly and easily: “instant software.” Taken to an extreme, it might become easier for a user to have an AI…
A $30,000 AI GPU doesn't outperform consumer GPUs at password cracking. Specops explains why attackers don't need exotic hardware to break weak passwords. Compute power is growing at an extraordinary pace.
Every year, millions of employees suffer illnesses and injuries that could have been prevented. As workplaces modernize, one thing remains constant: the need to protect our most valuable asset – people. Work shouldn’t come with a health…
Multi-extortion ransomware relies on stolen data to pressure victims with public leaks. Penta Security explains how its D.AMO platform keeps exfiltrated files encrypted and useless to attackers. the University of Mississippi Medical Cent…
Threat actors are exploiting the recent Claude Code source code leak by using fake GitHub repositories to deliver Vidar information-stealing malware.
A new malware-as-a-service called CrystalRAT is being promoted on Telegram, offering remote access, data theft, keylogging, and clipboard hijacking capabilities.
Anthropic confirmed that internal Claude Code source files were exposed through an npm packaging error, turning a release workflow mistake into a supply-chain and competitive-intelligence event.
In December, the Trump administration signed an executive order that neutered states’ ability to regulate AI by ordering his administration to both sue and withhold funds from states that try to do so. This action pointedly supported ind…
New governance development detected from EDPB News. Brussels, 23 March - On 17 March 2026, the EDPB conference “Cross-regulatory interplay and cooperation in the EU: a data protection perspective” took place in Brussels.
Japan’s election last month and the rise of the country’s newest and most innovative political party, Team Mirai, illustrates the viability of a different way to do politics. In this model, technology is used to make democratic processes…
Building confidence in sustainability claims worldwide. Sustainability claims are everywhere.
EU privacy regulators say the Digital Omnibus proposal can simplify compliance, but warn that narrowing the definition of personal data would weaken data protection.
The EDPB has signed a joint Global Privacy Assembly statement warning that AI-generated imagery and video can create serious privacy, dignity, and safety harms when real people are depicted without consent.
An EDPB update on cross-border data requests and AI training signals a more operational European privacy-governance agenda for organisations handling sensitive data.
A 2025 EDPB-backed report on LLM privacy risks focuses on concrete mitigations and real-world deployment scenarios rather than abstract AI principles.
A security and privacy watchlist covering student cyber safety policy, AI workplace tooling risk, and emerging privacy-governance signals worth monitoring.
Weak privacy controls can make AI systems harder to deploy, govern, and defend over time, turning privacy into an operational AI risk question.
AI security risk is increasingly concentrated in extensions, browser flows, workplace tooling, and other user-facing interfaces rather than models alone.