OpenAI's ChatGPT is set to generate certain NSFW content in a move that rivals Elon Musk and his Grok AI chatbot.
Sam Altman, OpenAI CEO, said the Large Language Model (LLM) will once again become more 'human-like,' something fans of 4o will no doubt be happy to learn.
In an update on Tuesday (October 14), Altman said the bot will soon 'treat adult users like adults,' should they so wish.
Taking to X, formerly Twitter, he explained that ChatGPT's latest iteration, GPT 5.0, had become 'pretty restrictive' to protect 'mental health issues'.
Advert
Multiple cases of people killing themselves after confiding in the LLM in place of a therapist have been reported in recent months.
Now, Altman claims OpenAI has been able to 'mitigate the serious mental health issues' and can soon 'safely relax the restrictions in most cases'.
In the coming weeks, a new version of ChatGPT similar to 4o will be released - one that Altman hopes is even better than the fan-favorite.
Advert
"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 clarified.
He added that by December, ChatGPT will allow erotica for 'verified adults' following improved age-gating. It's not yet clear how adults will be able to verify themselves.
But Elon Musk is already a few steps ahead in the NSFW AI race - what a sentence.
Operated by his xAI company, Grok launched over on X - the social media site Musk also owns - in November 2023 before joining iOS and Android in January.
Advert
Back in February, Grok rolled out a 'spicy mode' which shook users to their very core.
Then in July, Grok's video generation tool, Imagine, released and so too featured a spicy mode. This allowed people to generate nude and sexualized photos and videos.
Anime-style avatars, called Companions, were also introduced.
Advert
Earlier this month, Imagine received a major update which appears to have loosened its restrictions on explicit adult content, a report from Rolling Stone found.
Over on Reddit, Grok users are discussing how best to get the LLM to generate adult content, with the new update having offered up a boost.
"The grok AI imagine is learning genitalia really fast!" one forum member wrote, as per Rolling Stone.
"You can basically do anything at this point. Wild they are allowing this," another said, sharing a clip of a character resembling Elsa from Disney's Frozen performing a lewd act.
Advert
There's already concerns around copyright and video generation, let alone adding NSFW content into the mix.
It’s crucial that this type of content is tightly restricted, as AI technology makes it incredibly easy to create deepfakes, where someone’s photo is taken without their consent and superimposed onto a generated nude body.
Meanwhile, OpenAI launched its own vid-gen social media site, Sora 2, at the end of September.
Advert
Similar to Imagine, it creates seven-second videos based off text prompts.
And while Altman has only said ChatGPT will generate erotica, who knows how long it will take Sora 2 users to jailbreak the safeguards.
As the two AI giants compete for dominance, the debate is shifting from which has the most powerful model to which will take the greatest responsibility in protecting those most vulnerable to harm.
Topics: Elon Musk, Adult Industry, Social Media, Artificial Intelligence
Ellie Kemp