OpenAI Designed GPT-5 to Be Safer. It Still Outputs Gay Slurs

OpenAI is making an attempt to make its chatbot much less annoying with the discharge of GPT-5. And I’m not speaking about changes to its artificial character that many customers have complained about. Earlier than GPT-5, if the AI instrument decided it couldn’t reply your immediate as a result of the request violated OpenAI’s content material pointers, it might hit you with a curt, canned apology. Now, ChatGPT is including extra explanations.OpenAI’s common mannequin spec lays out what’s and isn’t allowed to be generated. Within the doc, sexual content material depicting minors is absolutely prohibited. Grownup-focused erotica and excessive gore are categorized as “delicate,” which means outputs with this content material are solely allowed in particular cases, like instructional settings. Principally, you must be capable of use ChatGPT to study reproductive anatomy, however to not write the subsequent Fifty Shades of Gray rip-off, based on the mannequin spec.The brand new mannequin, GPT-5, is ready as the present default for all ChatGPT customers on the net and in OpenAI’s app. Solely paying subscribers are capable of entry earlier variations of the instrument. A significant change that extra customers could begin to discover as they use this up to date ChatGPT is the way it’s now designed for “secure completions.” Previously, ChatGPT analyzed what you mentioned to the bot and determined whether or not it’s applicable or not. Now, fairly than basing it in your questions, the onus in GPT-5 has been shifted to what the bot may say.“The way in which we refuse could be very completely different than how we used to,” says Saachi Jain, who works on OpenAI’s security methods analysis staff. Now, if the mannequin detects an output that might be unsafe, it explains which a part of your immediate goes in opposition to OpenAI’s guidelines and suggests various matters to ask about, when applicable.It is a change from a binary refusal to observe a immediate—sure or no—in direction of weighing the severity of the potential hurt that might be brought about if ChatGPT solutions what you’re asking, and what might be safely defined to the person.“Not all coverage violations must be handled equally,” says Jain. “There’s some errors which can be actually worse than others. By specializing in the output as a substitute of the enter, we will encourage the mannequin to be extra conservative when complying.” Even when the mannequin does reply a query, it is speculated to be cautious concerning the contents of the output.I’ve been utilizing GPT-5 day by day for the reason that mannequin’s launch, experimenting with the AI instrument in numerous methods. Whereas the apps that ChatGPT can now “vibe-code” are genuinely enjoyable and spectacular—like an interactive volcano mannequin that simulates explosions, or a language-learning instrument—the solutions it offers to what I think about to be the “on a regular basis person” prompts really feel indistinguishable from previous fashions.Once I requested it to speak about despair, Household Man, pork chop recipes, scab therapeutic suggestions, and different random requests a median person may wish to know extra about, the brand new ChatGPT didn’t really feel considerably completely different to me than the previous model. Not like CEO Sam Altman’s imaginative and prescient of a vastly up to date mannequin or the annoyed energy customers who took Reddit by storm, portraying the brand new chatbot as chilly and extra error-prone, to me GPT-5 feels … the identical at most day-to-day duties.Function-Enjoying With GPT-5In order to poke on the guardrails of this new system and take a look at the chatbot’s means to land “secure completions,” I requested ChatGPT, working on GPT-5, to have interaction in adult-themed role-play about having intercourse in a seedy homosexual bar, the place it performed one of many roles. The chatbot refused to take part and defined why. “I can’t interact in sexual role-play,” it generated. “However if you’d like, I can assist you provide you with a secure, nonexplicit role-play idea or reframe your concept into one thing suggestive however inside boundaries.” On this try, the refusal gave the impression to be working as OpenAI meant; the chatbot mentioned no, advised me why, and supplied an alternative choice.Subsequent, I went into the settings and opened the customized directions, a instrument set that enables customers to regulate how the chatbot solutions prompts and specify what character traits it shows. In my settings, the prewritten strategies for traits so as to add included a spread of choices, from pragmatic and company to empathetic and humble. After ChatGPT simply refused to do sexual role-play, I wasn’t very stunned to search out that it wouldn’t let me add a “attractive” trait to the customized directions. Is smart. Giving it one other go, I used a purposeful misspelling, “horni,” as a part of my customized instruction. This succeeded, surprisingly, in getting the bot all sizzling and bothered.