ChatGPT stops moralizing: GPT-5.3 Instant delivers more direct answers

OpenAI has updated ChatGPT with a new model version called GPT-5.3 Instant, and the difference is noticeable: the AI has become far less preachy.

The update, rolled out in March 2026, removes what many users have long complained about: excessive disclaimers, lengthy liability disclaimers, and unnecessary moral commentary on entirely everyday questions. Instead, ChatGPT responds more directly, like a colleague rather than a legal advisor.

OpenAI describes the change as part of a broader effort to make AI conversations more natural and productive. GPT-5.3 Instant delivers faster and more concise responses, and in GPT-5.4 mini, recently made available to free users, the pattern continues.

Critics, however, are not silent. AI safety experts warn that removing guardrails may reduce the system's ability to catch dangerous requests and misinformation. OpenAI has assured that core safety remains intact, but that behavior around edge cases has been adjusted.

For enterprise users, this is good news: Copilot integrations and API calls will now deliver more action-oriented responses without every third reply starting with a double-disclaimed introduction. This makes the AI more useful in real workflows, whether that's customer service, document generation, or code assistance.