OpenAI Publishes June 2026 Threat Report Detailing Malicious Uses of AI

OpenAI released its June 2026 Report on Malicious Uses of AI, a publicly available PDF documenting threat actors' attempts to exploit the company's models for harmful purposes. The report, hosted directly on OpenAI's content delivery network, covers categories of misuse the lab has observed and intervened against across its platform. Specific findings and case studies are detailed within the document, which OpenAI has made available without a paywall. The release continues a pattern of periodic transparency reporting from the lab on how its systems are being weaponized or probed by bad actors. No embargo date or accompanying press release has been identified at time of publication. Separately, Somantix Ledger reported that OpenAI is laying groundwork for an on-premises enterprise product; OpenAI has not officially confirmed this, and the report should be treated as unconfirmed. The threat report is available at the OpenAI CDN link referenced in the company's documentation.