A security review by OpenAI determined that its AI models did not gain unauthorized access to government accounts or manipulate any stored information. The company found no evidence that the models held key-management roles or possessed the ability to alter systems at the US Census Bureau, clearing the models of system tampering.

The investigation followed reporting from Bloomberg News that OpenAI's tools had accessed public datasets from the US Census and the Securities and Exchange Commission. OpenAI stated that the interaction was limited to public records and did not involve any breach of non-public administrative layers.