OpenAI is working with rivals Anthropic and Google DeepMind on artificial intelligence safety, according to a Bloomberg News report published on September 15, 2026. OpenAI Global Policy Chief Chris Lehane confirmed that the companies have been working together on safety issues for several weeks. The discussions involve companies developing some of the world's most advanced AI systems and focus on managing emerging risks. The cooperation comes as concerns about the safety of increasingly capable AI models continue to grow. The talks show that competing AI companies are also exploring areas where common safety approaches may be useful.Why Are AI Rivals Working Together?The main reason for the discussions is to address AI safety challenges that can affect the wider technology industry. OpenAI, Anthropic and Google DeepMind compete directly in developing advanced AI models, but their systems face some similar technical and security risks. Lehane said the companies have been holding discussions for weeks, while reports indicate the talks are focused on safety rather than commercial cooperation. The development comes as governments, researchers and technology companies debate how advanced AI should be tested and controlled.OpenAI is discussing AI safety with Anthropic and Google DeepMind.The discussions have reportedly continued for several weeks.OpenAI Global Policy Chief Chris Lehane confirmed the cooperation.The companies remain competitors in the AI market.The discussions are focused on safety-related issues rather than a business merger or partnership.Growing Focus On AI SafetyAI safety has become a major issue as frontier models gain stronger capabilities in coding, cybersecurity, research and autonomous task handling. OpenAI has recently published work on long-running AI systems and said new evaluations were needed after identifying failures that existing tests did not capture. The company has also said that increasingly capable systems require stronger safety and security measures. Anthropic similarly maintains a Responsible Scaling Policy that links stronger safeguards to rising model capabilities and risks.The companies have already explored cooperation on safety research in the past. In 2025, OpenAI and Anthropic conducted a joint evaluation in which each company tested the other's publicly released models for alignment and safety issues. OpenAI said the exercise was intended to identify gaps that might be missed by internal testing alone. The latest discussions expand the reported cooperation to include Google DeepMind.Also Read: Anthropic: Claude AI Used For Weapons Software Before Failed Test ReviewLatest AI Safety DiscussionsThe latest development is significant because it brings three major AI developers into safety discussions while they continue competing over models and products. Reuters, citing the Bloomberg report, said Lehane confirmed the discussions and reported that OpenAI does not believe an antitrust waiver is required for its current safety work with the other companies. The discussions come at a time when concerns over advanced AI risks are receiving greater attention from companies and policymakers. OpenAI has separately called for mandatory national AI safety requirements and stronger safety measures for advanced systems.Key HighlightsOpenAI is working with Anthropic and Google DeepMind on AI safety.The discussions have reportedly been taking place for several weeks.Chris Lehane confirmed the cooperation, according to the reported statements.The three companies remain competitors in the AI industry.OpenAI and Anthropic have previously collaborated on model safety evaluations.The latest talks come amid wider discussions about frontier AI risks and safeguards.What Comes Next For AI SafetyThe reported cooperation does not mean the three companies have created a single AI safety organization or adopted one common safety framework. Instead, the available reporting points to discussions between competing companies about shared safety concerns and possible areas of cooperation. Each company continues to maintain its own safety policies, evaluation methods and governance frameworks. The future impact of these talks will depend on whether the discussions lead to shared standards, testing methods or other practical safety measures.The latest discussions between OpenAI, Anthropic and Google DeepMind highlight the growing importance of AI safety as advanced models become more capable. The companies continue to compete commercially, while also discussing risks that can affect the broader AI ecosystem. For now, the reported cooperation remains focused on safety discussions rather than a broader business partnership. Further details will be needed to determine whether the talks result in formal joint standards or additional industry-wide safety efforts.Disclaimer: This content is provided only for knowledge purposes and to provide general news information; readers should refer to official announcements for future updates.
