OpenAI will watermark ChatGPT outputs by default—but only in the EU

OpenAI is making ChatGPT text watermarking the default in the EU to align with the EU AI Act, signaling that AI governance is moving from policy discussion to enforceable product requirements. For CIOs and technology leaders, this means AI adoption and content workflows will increasingly need region-aware controls, auditability, and compliance processes, while also recognizing that watermarking remains imperfect and can be degraded or bypassed. IT organizations should expect a growing need to balance productivity gains from generative AI with legal, reputational, and operational risk management across jurisdictions.