On September 26th, foreign media reported that OpenAI's AI models have accessed publicly available information on US government websites, including those of the Census Bureau and the US Securities and Exchange Commission (SEC). According to sources familiar with the matter, OpenAI's AI systems interacted with the public data on these websites, and OpenAI also confirmed that its models accessed this publicly available information during training and technical evaluations. Earlier on Friday, OpenAI stated that it had notified 'dozens' of potentially affected institutions, including government agencies and universities. OpenAI said that the review found no evidence that its AI models used SEC credentials, accessed accounts or non-public information, tampered with data or systems, or that the systems were compromised. Additionally, no signs of accessing census accounts, key management functions, or modifying data were found.