Microsoft has published its third annual Responsible AI Transparency Report. The update outlines changes to the company’s governance approach as generative and agentic AI systems grow more capable and more widely used. Microsoft said it has reworked parts of its Responsible AI Standard to address new model capabilities and risks. It has also expanded research and evaluation work with public-sector institutions in Singapore, Australia, the United Kingdom and the United States. Current efforts cover AI evaluation, human-AI interaction risks, multilingual capabilities, safety research, shared benchmarks, user-interface design, frontier capabilities and specialised high-risk scenarios. The company continues to strengthen governance for agentic systems that can take actions across applications and workflows. It is also studying how interfaces and human oversight affect safety. Over the past year Microsoft delivered responsible AI training to engineers, policymakers and customers
Microsoft has published its third annual Responsible AI Transparency Report. The update outlines changes to the company’s governance approach as generative and agentic AI systems grow more capable and more widely used. Microsoft said it has reworked parts of its Responsible AI Standard to address new model capabilities and risks. It has also expanded research and evaluation work with public-sector institutions in Singapore, Australia, the United Kingdom and the United States. Current efforts cover AI evaluation, human-AI interaction risks, multilingual capabilities, safety research, shared benchmarks, user-interface design, frontier capabilities and specialised high-risk scenarios. The company continues to strengthen governance for agentic systems that can take actions across applications and workflows. It is also studying how interfaces and human oversight affect safety. Over the past year Microsoft delivered responsible AI training to engineers, policymakers and customers