Lord Forgive Me, It’s Time to Go Back to the Old ChatGPT

Earlier this year, OpenAI scaled back some of ChatGPT’s “personality” as part of a broader effort to improve user safety following the death of a teenager who took his own life after discussing it with the chatbot. But apparently, that’s all in the past. Sam Altman announced on Twitter that the company is going back to the old ChatGPT, now with porn mode.

“We made ChatGPT pretty restrictive to make sure we were being careful with mental health issues,” Altman said, referring to the company’s age-gating that pushed users into a more age-appropriate experience. Around the same time, users started complaining about ChatGPT getting “lobotomized,” providing worse outputs and less personality.  “We realize this made it less useful/enjoyable to many users who had no mental health problems, but given the seriousness of the issue we wanted to get this right.” That change followed the filing of a wrongful death lawsuit from the parents of a 16-year-old who asked ChatGPT, among other things, for advice on how to tie a noose before taking his own life.

But don’t worry, that’s all fixed now! Despite admitting earlier this year that safeguards can “degrade” over the course of longer conversations, Altman confidently claimed, “We have been able to mitigate the serious mental health issues.” Because of that, the company believes it can “safely relax the restrictions in most cases.” In the coming weeks, according to Altman, ChatGPT will be allowed to have more of a personality, like the company’s previous 4o model. When the company upgraded its model to GPT-5 earlier this year, users began grieving the loss of their AI companion and lamenting the chatbot’s more sterile responses. You know, just regular healthy behaviors.

“If you want your ChatGPT to respond in a very human-like way, or use a ton of emoji, or act like a friend, ChatGPT should do it (but only if you want it, not because we are usage-maxxing),” Altman said, apparently ignoring the company’s own previous reporting that warned people could develop an “emotional reliance” when interacting with its 4o model. MIT researchers have warned that users who “perceive or desire an AI to have caring motives will use language that elicits precisely this behavior. This creates an echo chamber of affection that threatens to be extremely addictive.” Now that’s apparently a feature and not a bug. Very cool.

Taking it a step further, Altman said the company would further embrace its “treat adult users like adults” principle by introducing “erotica for verified adults.” Earlier this year, Altman mocked Elon Musk’s xAI for releasing an AI girlfriend mode. Turns out he’s come around on the waifu way.