Simon Willison45 · 4d ago
llm 0.31
Release: llm 0.31 New GPT-5.5 OpenAI model: llm -m gpt-5.5 . #1418 New option to set the text verbosity level for GPT-5+ OpenAI models: -o verbosity low . Values are low , medium , high . New option for setting the image detail level used for image attachments to OpenAI models: -o image_detail low - values are low , high and auto , and GPT-5.4 and 5.5 also accept original . Models listed in extra-openai-models.yaml are now also registered as asynchronous. #1395 Tags: gpt , openai , llm
Security Boulevard32 · 4d ago
Benchmarking OpenAI’s Privacy Filter: What it gets right, and where PII detection still needs real data
A deep dive into OpenAI’s Privacy Filter, benchmarking its PII detection performance against Tonic Textual on real-world datasets. We explore where the model succeeds, where it struggles, and how fine-tuning with labeled data impacts accuracy across domains like healthcare, legal, and web data. The post Benchmarking OpenAI’s Privacy Filter: What it gets right, and where PII detection still needs real data appeared first on Security Boulevard .
The Hacker News28 · 5d ago
LMDeploy CVE-2026-33626 Flaw Exploited Within 13 Hours of Disclosure
A high-severity security flaw in LMDeploy, an open-source toolkit for compressing, deploying, and serving LLMs, has come under active exploitation in the wild less than 13 hours after its public disclosure. The vulnerability, tracked as CVE-2026-33626 (CVSS score: 7.5), relates to a Server-Side Request Forgery (SSRF) vulnerability that could be exploited to access sensitive data. "A server-side
The Hacker News25 · 5d ago
NASA Employees Duped in Chinese Phishing Scheme Targeting U.S. Defense Software
The Office of Inspector General (OIG) of the U.S. National Aeronautics and Space Administration (NASA) has revealed how a Chinese national posed as a U.S. researcher as part of a spear-phishing campaign to obtain sensitive information from the space agency, as well as from government entities, universities, and private companies, in violation of export control laws. "For years, NASA employees
Latest stories for ZDNET in Security20 · 4d ago
How I used Claude AI to plan an entire hiking trip to the Adirondacks in 30 minutes - for free
Using Claude's interactive connections to third-party services such as TripAdvisor and AllTrails, I mapped a summer hiking trip, including trails, hotels, tours, and even a playlist to accompany us.
The Register (Security)20 · 5d ago
Intel bets the farm on AI inference to drag CPU back to the top table
Chipzilla hopes agents, robots, and edge devices make CPUs cool again... now it has to build the chips Intel is betting on AI to reverse its fortunes, wagering that inference and agentic workloads will restore the CPU to the center of compute - even as its chip manufacturing struggles persist.…
Security on TechRepublic17 · 4d ago
Health Records of 500,000 UK Biobank Volunteers Listed Online in China
Health data from 500,000 UK Biobank participants was found listed for sale online in China, raising concerns over research access misuse and data security. The post Health Records of 500,000 UK Biobank Volunteers Listed Online in China appeared first on TechRepublic .
Security Boulevard17 · 5d ago
[un]prompted 2026 – Building Secure Agentic Systems: Lessons From Daily-Driver Agents
Author, Creator & Presenter: Brooks McMillin, AI Security Researcher & Security Engineer, Dropbox Our thanks to [un]prompted for publishing their Creators, Authors and Presenter’s outstanding [un]prompted 2026 AI Security Practitioner content on the Organizations' YouTube Channel . Permalink The post [un]prompted 2026 – Building Secure Agentic Systems: Lessons From Daily-Driver Agents appeared first on Security Boulevard .
The Register (Security)17 · 5d ago
It's a myth that you need Mythos to find bugs: Open source models can do it just as well
OpenAI's first security hire, Ari Herbert-Voss, thinks more automated bug finding will improve security without costing jobs Black Hat Asia Open source models can find bugs as effectively as Anthropic's Mythos, according to Ari Herbert-Voss, CEO of AI-powered security startup RunSybil and OpenAI's first security hire.…
The Hacker News13 · 5d ago
Bridging the AI Agent Authority Gap: Continuous Observability as the Decision Engine
The AI Agent Authority Gap - From Ungoverned to Delegation As discussed in our previous article, AI agents are exposing a structural gap in enterprise security, but the problem is often framed too narrowly. The issue is not simply that agents are new actors. It is that agents are delegated actors. They do not emerge with independent authority. They are triggered, invoked, provisioned, or
IT Security Guru 11 · 5d ago
NCSC chief warns of ‘perfect storm’ as cyber threats intensify at CyberUK
At this week’s CyberUK conference in Glasgow, National Cyber Security Centre (NCSC) CEO Richard Horne delivered a stark assessment of the evolving cyber threat landscape, warning that organisations are facing a “perfect storm” driven by rapid advances in artificial intelligence and rising geopolitical tensions. In his keynote, Horne highlighted how emerging technologies are fundamentally reshaping […] The post NCSC chief warns of ‘perfect storm’ as cyber threats intensify at CyberUK appeared ...
Cyber Kendra11 · 5d ago
Your TLS Certificates Are About to Silently Break — And You Won't Know Until It's Too Late
A quiet but consequential change is coming to internet security infrastructure — and unlike most breaking changes, this one won't announce itself until it's already caused an outage. Starting May 2026 , major public Certificate Authorities, including Let's Encrypt, DigiCert, Sectigo, and GlobalSign, will stop embedding the Client Authentication Extended Key Usage (EKU) flag in their publicly issued TLS certificates. The CA/Browser Forum — the governing body that sets browser and CA policy — ...
MIT Technology Review8 · 5d ago
Health-care AI is here. We don’t know if it actually helps patients.
I don’t need to tell you that AI is everywhere. Or that it is being used, increasingly, in hospitals. Doctors are using AI to help them with note-taking. AI-based tools are trawling through patient records, flagging people who may require certain support or treatments. They are also used to interpret medical exam results and x-rays. A…