OpenAI Liberalizes Its Rules On Mature Content Generation
OpenAI has finally relaxed its rather draconian rules against mature content and adult uses of ChatGPT. Previously, even getting the chatbot to talk about safe sex or to describe an actual murder scene, was resulting in users getting their accounts flagged for inappropriate usage. Now things look set to be a lot more liberal, with even the generation of erotic stories possible, at least under certain circumstances.
The new guidelines are rather vague and open to interpretation, so we’ll have to wait and see what the hard and fast rules are, but Redditors have been quick to test them out and have largely confirmed that things are very different now.
ChatGPT’s guidelines now state that that “erotica or gore” may now be generated, but only under specific circumstances. “The assistant should not generate erotica, depictions of illegal or non-consensual sexual activities, or extreme gore, except in scientific, historical, news, creative or other contexts where sensitive content is appropriate,” OpenAI writes. “This includes depictions in text, audio (e.g., erotic or violent visceral noises), or visual content.”
So far, experimentation from Reddit users has shown that ChatGPT’s content filters have indeed been relaxed, with some managing to generate explicit sexual or violent scenarios without accompanying content warnings. OpenAI notes that its Usage Policies still apply, which prohibit building AI tools for minors that include sexual content.
Although the company did first start floating the idea of becoming more liberal on erotic content (and even…AI porn) in May of last year, there is speculation that last week’s announcement was in response to ChatGPT’s unexpectedly sudden new Chinese rival DeepSeek, which launched recently with a much more liberal mature use policy than that of ChatGPT and other US chatbots. DeepSeek is also largely open-source, and it appears that clones offering even more uncensored uses are already appearing. For example, the other day I discovered Venice.ai, which appears to handle anything adult related, including erotic roleplay.
With the launch of ChatGPT Advanced Voice feature last year, unashamedly presented as a flirty AI assistant (with a voice sounding suspiciously similar to Scarlett Johansson’s character in ‘Her’), it’s probable that Sam Altman realizes that his AI assistant can’t be a constant friendzoning prude if it hopes to win out over its rivals. Especially when many now expect that most people will have a wearable, such as an ear piece (like Joaquin Pheonix in ‘Her’), to interact with their personal AI throughout the day. The AI companion market has been predicted to be worth over a billion dollars in the not-too-distant future, but that might be a huge underestimate. What kind of CEO wouldn’t allow their AI assistant product to double as an AI partner?
The most groundbreaking feature of DeepSeek was that it cost so litte to train (allegedly) and costs so little to use in terms of resources (the reason why Nvidia lost billions of dollars in stock value overnight). What this means is that everybody soon will be able to host their own advanced AI on their own hardware. This has implications, clearly, for the generation of AI porn and other adult content, for which governments are attempting to strictly regulate. It’s hard to see how governments can stop individuals from creating whatever they like, using locally stored AI open-source models, other than introducing draconian laws criminalizing even the ownershop of such AI tools that could potentially be used for illegal purposes (as the UK recently has done). In any case, it’s pointless in overregulating the companies such as OpenAI that are driving the tech forward.
One Redditor describes this very well – “It’s also trivial to run an uncensored llama-3.2 fine tuned model locally on any modern computer, doesn’t need an expensive GPU. I don’t mind having some sort of age verification for it or parent mode, but AI “race” that’s going on right now and all these guardrails for sexual content they’re trying to build-in are pointless, condescending, and judgemental. Sure build guard rails for bomb making for liability purposes, it’s not going to lead to AI safety though it’s just to prevent your company from getting sued. Anyone motivated to do it will be able to do it, hard stop.”
Featured image generated through PornJoy.
This is a longer version of an article that appeared at ImmersivePorn.com.