OpenAI unlocks 'adult mode'. ChatGpt increasingly 'human'?

"If you want your Chatpt to respond very much like a human, or use a lot of emoji, or act like a friend, ChatGPT should do it (but only if you want it to, not because we are maximising usage)."

"In December, with the introduction of a more comprehensive age-gating system and in line with our principle of 'treating adult users as adults', we will allow even more, such as verified adult erotica."

We don't know how far it will go down the sex chat path. It could be something similar to what Elon Musk's Grok already does, i.e. soft allusive porn, erotic role playing and partial graphic nudity. Grok has introduced the companions function (virtual companions) with characters like Ani, a sensual anime designed for romantic or erotic interactions. Ani can 'wear lingerie', cast seductive glances and establish dialogues that push towards an affective-sexual dimension.

Grok Imagine allows the generation of images and short videos (animations) from textual prompts or reference images. In this tool there is a 'spicy' mode that allows the creation of partial sexual content: controlled nudity, less explicit erotic content, sexualised images

Why the turnaround? It is not surprising: it is along the lines of what Meta has also done, which has toned down moderation since an anti-woke president like Donald Trump and allowed erotic conversations in its AI chats. The idea is that big tech can now push engagement without risking regulatory shingles, at least in the most important market, the US.

For OpenAI, the business motivation is obvious. After gaining over 800 million weekly users, it must find new ways to increase the platform's usage time and consolidate loyalty. Also in view of the need, sooner or later, to cut its losses to zero (not before 2029, according to Altman).

Erotic or romantic interaction is, in terms of retention, a powerful lever for these purposes. A strategy on which Character.AI has always relied, where millions of users spend hours in intimate or fantasy conversations with customised chatbots. Altman knows that every extra minute spent on ChatGpt is a value multiplier: more data, more subscriptions, more revenue.

The move, however, opens up an unprecedented risk front. On the technical side, the reliability of age verification systems is far from guaranteed. Altman anticipated that OpenAI will be based on an 'age prediction' mechanism, capable of estimating the user's age from usage patterns and, in case of error, will require the user to submit an identity document. Europe has launched an age verification system with certified third parties, which is more secure and more privacy-protecting than the exchange of documents https://www.ilsole24ore.com/art/ecco-come-cambia-l-accesso-social-gli-adolescenti-AHP1TzyB. It remains to be seen to what extent Chatgpt's breakthrough will reach Europe and how it will be verified.

The social and psychological risks have to be considered. Some vulnerable users have already developed illusory or pathological relationships with AI. There are also cases of young people who have ended up committing suicide. Ongoing lawsuits in the US, filed against OpenAI by the parents of the young people who lost their lives, will have to ascertain the responsibilities of the tech company. With the debut of Eros, the situation could become even more complicated.