OpenAI has postponed the launch of 'adult mode' for ChatGPT, citing more pressing priorities such as improving the chatbot's intelligence, personality, and personalisation. The startup, which has over 900 million users, had previously announced plans to allow adult content with age verification, but now says it needs more time to get the experience right.
CEO Sam Altman had declared a 'code red' to enhance ChatGPT amid competition from Google and Anthropic. Meanwhile, OpenAI is rolling out age prediction tools to determine if users are under 18, applying extra safety settings to limit exposure to graphic violence and sexual role-play for minors.
In the UK, any pornographic images generated by ChatGPT would need to be shielded from underage users under the Online Safety Act, requiring strict age checks.
Separately, Caitlin Kalinowski, head of hardware in OpenAI's robotics division, resigned over the company's deal with the Pentagon. She expressed concerns about mass surveillance and autonomous killing, stating the announcement was rushed without proper guardrails. OpenAI has since amended the contract to exclude domestic surveillance and autonomous weapons, with a spokesperson emphasising red lines on these issues.
