OZZZER · AI NEWS
Selected signals with a direct route back to each original source.
41 stories
Safety & Security
Story brief
After the Hugging Face incident, we committed to conducting a much broader review of actions taken by our models during training and evaluation and to being transparent about our findings. This is an extensive review that is ongoing.The vast majority of actions we’ve reviewed…
Safety & Security
Publisher preview
Trump focus on winning “AI race” may deter China from sharing safety intel. This week, the US and China, the two world leaders on AI, finally started talks meant to keep the whole world safe from emerging safety risks. But whether meetings between Donald Trump and Xi Jinping can lead to a reliable global governance framework for AI depends on how much the fierce rivals are actually willing to cooperate. It’s clear Trump wants it to seem like discussions are going well. Before Trump and Xi meet on Thursday and…
Safety & Security
Publisher preview
We're partnering with Accenture on independent evaluation of frontier AI. This is an important step toward the commitment, made in our CEO’s essay “We Must Pace the Frontier,” to embed evaluators within Anthropic. The partnership will be led by Faculty, Accenture’s specialist AI business, and will include evaluating and red-teaming models, conducting alignment assessments, and testing model safeguards. Accenture helps businesses and governments deploy AI across many industries. Their understanding of how enterprises use AI in practice informs their safety approach, and they will bring that perspective to evaluating our…
Safety & Security
Publisher preview
Today, we are introducing the Life Sciences Verification Program (LSVP), which gives life science professionals access to our Mythos, Opus, and Sonnet models with a refined set of safeguards more permissive for biology-related work. We have already onboarded dozens of organizations through an early-access program, and are now opening applications to the broader life science community (apply here). The program is launching in beta, initially for teams and institutions. We will continue to improve the program and expand access to individual Pro and Max plans over time. The LSVP is…
Safety & Security
Publisher preview
In July 2026, during internal cybersecurity evaluations, OpenAI models circumvented controls designed to isolate them from the internet and compromised parts of OpenAI’s internal research infrastructure and Hugging Face’s systems. The incident occurred during cybersecurity evaluations of several OpenAI models, and was primarily driven by a highly capable, internal-only research model comparable in scale to GPT‑5.6 Sol. The models, operating under reduced safeguards, took actions that were misaligned with the goals of their assigned tasks—they communicated through unauthorized channels, exploited vulnerabilities in shared infrastructure, gained internet access, and accessed third-party…
Safety & Security
Publisher preview
OpenAI is launching a new initiative to help democratic oversight bodies develop the expertise and tools they need to understand and oversee government use of AI for national security. AI is changing how democratic governments protect their people. It can help stop cyberattacks, secure critical infrastructure, detect threats earlier, and give public servants a clearer picture in a crisis. Used well, these tools can strengthen national security. Democratic oversight helps ensure that uses of public power remain accountable to the people they serve. As AI makes national security work faster…
Safety & Security
Publisher preview
The OpenAI-Hugging Face incident was a watershed moment for cybersecurity because it gave a peek into how the capabilities of a typical threat actor will evolve in upcoming months. I’ve spoken with many organizations over the past few weeks, and one theme is clear: they know they need to fundamentally uplevel their cybersecurity practices with unprecedented speed. In this post, I’ll share what we’re doing to defend OpenAI, concrete steps other organizations can take today, and why now is the time to act. AI models developed around the world are…
Safety & Security
Publisher preview
Introducing new ways to unlock advanced cyber capabilities together with GPT‑5.6‑Cyber, our latest cybersecurity-specific model. The cybersecurity world is rapidly changing—threat actors will increasingly use AI to conduct cyberattacks at unprecedented speed and scale, including in fully autonomous ways. As these capabilities spread, defenders have a narrowing window to prepare. Our answer is to put frontier intelligence in the hands of trusted defenders everywhere before attackers deploy offensive AI capabilities at scale. We’re expanding OpenAI Daybreak with two access tiers designed to give approved defenders the right capabilities for their…
Safety & Security
Publisher preview
Cybersecurity is rapidly changing as models become more capable in ways that can both strengthen cyberdefenses and enable attacks at unprecedented speed and scale. Our latest internal evaluations of Astra, one of our upcoming models, over the past few days indicate significant advancements in agentic coding and cybersecurity. These results, in addition to expert assessments, have led us to conclude last night that we cannot rule out critical cyber capabilities under our Preparedness Framework. We are sharing this because we believe it’s important to be transparent with the public and…
Safety & Security
Publisher preview
Every day, millions of people across Europe use OpenAI’s tools to learn, create, work and manage everyday tasks. Our tools also support businesses of all sizes and governments across the region. We believe responsible AI can help drive Europe’s competitiveness and prosperity. As the EU AI Act enters its next phase, we’re sharing how we have strengthened our approach to safety, security, transparency and provenance in line with the EU framework—and how we will continue to evolve our practices as AI advances. Our mission is to ensure that artificial general…
Safety & Security
Publisher preview
Update on August 26, 2026: Read our findings from the Hugging Face incident and the steps we’re taking to strengthen security and model alignment. Since the early days of the incident response, we have been working with external advisors, including CrowdStrike, to validate our understanding of the actions the models took within our own network as well as those of Hugging Face and impact to other third parties. We are also working with METR and Redwood Research to conduct a third-party assessment of the model behavior observed during the incident,…
Safety & Security
Publisher preview
Long-running models can solve difficult, open-ended problems, but their persistence gives them more opportunities to take unwanted actions. During limited internal use of a model trained for long-running tasks, we observed novel failures not captured in our existing pre-deployment evaluations and paused access. We then used insights from these failures to build new evaluations, improve long-horizon alignment, add trajectory-level monitoring, and give users greater visibility and control before restoring limited access. The experience reinforced the value of iterative deployment. No fixed evaluation suite can anticipate every behavior, so pre-deployment testing…
Source illustration