The archive
Regulation, standards, safety research and privacy.
OpenAI says it will give Ukraine's government access to Daybreak, its AI program for security defenders, extending a push that already covers ENISA, CERT Polska and others in Europe.
5 min read
OpenAI's CEO used a Security Council appearance to call for shared frontier AI standards, incident reporting and human oversight — and to claim a model solved a Millennium Prize Problem.
5 min read
OpenAI says more than 80 licensed clinicians in 22 countries wrote the rubrics for MentalHealthBench, an open evaluation of AI responses across everyday stress, high distress and emergencies.
5 min read
No single filter makes an AI application safe. This guide explains how to layer input checks, model instructions, output validation, permissions, and monitoring into a practical defense.
8 min read
OpenAI has laid out four priority areas and seven principles for independent safety assessments of its models. The document is detailed on scope and access, quiet on who pays and who publishes.
5 min read
OpenAI claims an internal model resolved the Navier–Stokes Millennium Prize problem and 100-plus open problems, and is convening an independent mathematics advisory group at IAS.
5 min read
OpenAI published a proposal urging the United States to lead the creation of international technical standards for frontier AI, including for recursive self-improvement.
5 min read
AI features create new ways for personal data to leak, persist, or be misused. These are the core privacy practices teams should build in from the first prototype.
7 min read
OpenAI released a youth safety blueprint aimed at Australian policymakers, setting out six pillars covering AI literacy, age assurance, crisis support and parental controls.
5 min read
Mila and Mozilla announced an initiative to build an open source AI foundation layer, with $5 million from Mozilla and $1 million from Hypertec for Canadian deployments.
5 min read
Mila and Germany's DFKI signed a memorandum of understanding at ALL IN in Montreal, committing to joint research on energy applications, AI safety and agentic AI.
5 min read
AI bias rarely comes from a single bad decision. It builds up across data, labels, design choices, and deployment. This guide traces each source and the practical steps teams use to reduce it.
8 min read