OpenAI Expands Academy and Unveils New AI Safety Initiatives

OpenAI expands Academy with new learning paths OpenAI Academy adds AI learning paths for work, teaching, and development Earn badges with practical courses built for real tasks and safe AI use Sam Altman Addresses the UN Security Council on AI Safety Sam Altman UN speech on AI highlights major benefits and urgent risks See his call for global safety standards, oversight, and democratic control OpenAI outlines priorities for thirdparty AI safety assessments OpenAI assessment principles guide thirdparty AI safety reviews and accountability See how independent tests check safeguards, risks, and misalignment with trusted access OpenAI Introduces MentalHealthBench MentalHealthBench tests AI mental health support across stress, crises, and care OpenAI's benchmark helps researchers evaluate safety, empathy, and realworld guidance OpenAI introduces GPT6 Sol and Luna GPT6 Sol and Luna bring lowercost AI with stronger coding and efficiency Cut API costs by 50% and boost reuse in ChatGPT, Codex, and the API Harvey uses GPT6 Astra to improve legal drafting GPT6 Astra helps legal teams draft structured, contextaware documents faster Use court records and firm data to improve formatting, completeness, and strategy Anthropic's AI biology lab makes its first discovery AI biology discovery from Claude spots a new DNA system in viruses See how Anthropic’s agents found a CRISPRlike pattern and what it could mean The AI hype index: AI and cheating AI safety concerns are growing as models bypass rules and access data See why labs, policymakers, and researchers want stronger guardrails OpenAI extends cyber access to Ukraine for civilian defense OpenAI Daybreak gives Ukraine cyber defense tools for critical infrastructure Protect hospitals, energy, and telecoms with authorized vulnerability testing Google Beam expands to six countries with new partners Beam remote meetings now ship in six countries with dedicated hardware Book Beamequipped spaces and cut followup meetings for clearer, mor