OpenAI has once again postponed the rollout of an “adult mode” for ChatGPT, a feature that was expected to give verified adult users access to erotica and other NSFW content. Sam Altman originally said the feature would ship in December 2025, framing it as a “treat adults like adults” principle. The launch was first pushed to the first quarter of 2026 after an internal “code red,” and OpenAI has now delayed it again without a new date.
In its comment to Axios, OpenAI said its current focus is on more foundational priorities, including improving the model’s intelligence and pushing further on personalization and proactive behavior. The subtext is pretty clear: shipping a robust age gate that regulators will actually accept is a heavy lift, and right now the company is prioritizing core product performance and user experience over a niche mode that brings reputational and compliance risk.
The delay also lands in a moment when age assurance is becoming more explicit and enforceable in multiple jurisdictions. That matters because “adult mode” is not just a toggle, it depends on reliable age verification or age estimation, and that infrastructure has its own privacy, security, and false positive problems. Reuters reported OpenAI rolling out an age prediction feature for ChatGPT, with additional protections for accounts estimated to belong to minors, and an option to restore full access through identity verification via an external provider. This kind of system is exactly the prerequisite OpenAI has been pointing to, and it is also exactly the kind of system that draws scrutiny.
There is also a product strategy angle. “Adult mode” is easy to announce and hard to ship safely at scale, especially when the platform is under pressure to keep improving quality and keep up with competitors. Axios and TechCrunch both describe the shift as a prioritization call rather than a cancellation: OpenAI is still signaling the idea remains on the roadmap, but it is not the workstream they want to staff aggressively while they are tightening the mainline ChatGPT experience.
For users who were waiting for looser content filters, the practical implication is that the timeline is now open-ended. Even if OpenAI keeps moving toward broader “adult” allowances, expect it to be tightly coupled to age assurance, jurisdiction-specific rules, and conservative rollout steps rather than a big, global switch-flip. The legal and regulatory climate around minors, harmful content, and verification is moving in the opposite direction of “ship fast,” which makes indefinite delays a rational outcome even if the feature remains conceptually alive.
