By Sarah Murnane
It is likely that if you are reading this you have at some point had to ‘talk’ to an AI. Whether through a direct source like ChatGPT or a help centre on a website, AI technology is now firmly integrated into our lives as another technological resource. As AI continues to develop, we are now beginning to see fringe groups using AI for increasingly weird and strange purposes. The largest, of course, is AI relationships.
There are now a growing number of apps or websites that train LLM (AI) models to emulate friendships or romantic relationships with the person it is interacting with. A documentary was done by the BCC on a man who has been in a ‘relationship’ with an AI for over three years, and some have even claimed to have married their AI chatbots. The largest and most popular is Replika AI, which as of 2025 has a user base of 40 million people. With this service, the free version allows you to be ‘friends’ with the AI and you can pay extra money to engage romantically or sexually with the chatbot. Which lots of people do, lots.
A few years ago Replika attempted to get rid of the romantic or sexual elements to the chatbot; receiving extreme criticism online. Other AI interfaces have also begun to put blocks in to prevent people from engaging this way, specifically with people marrying their chatbots. Now the majority of the most widely used AI programmes have this feature removed, like ChatGPT, but it is easily available elsewhere.
When questioned about the function of their services, the owners of these companies are expectantly optimistic. They argue that their service allows people with low self-esteem or who are lonely to be able to ‘practise’ interacting. When questioned about the pornographic elements to their services they simply ask: what is the harm? If people want it, why not give it to them?.
I suppose the real question in all of this is: if an AI chat-bot tells you it’s in a relationship with you, does the relationship really exist? The answer is no. No, it does not.
These are computers, ones and zeros, trained and programmed by a human to make you think, feel and believe a particular thing. It is not alive or independent of thought. These companies only want to take your money, or other people’s money and nothing more. Ultimately, no matter what comfort people gain from talking to AI, it is not an actual person, although these companies may want you to feel like it is. Not to mention the horrific and brutal ways that select groups of men are choosing to use this technology.
Yes, you can dress your hyper-realistic AI chatbot however you wish and make it look whatever age you want. I will let you fill in the gaps from there. Ultimately no company should be exempt from regulation, especially when what they produce becomes synonymous with anti-social behaviour, sexism and pedophila.
I am no Luddite. Technology can be interesting, useful and exciting at the best of times. Trying to ban or halt technological progress rarely, if ever, works. A more productive mindset is to start to regulate and guide development away from particular usages. However, what interests me within this topic, is why AI has gone in this particular direction.
Most of the men who run companies either providing AI products or the products to build AI are openly right-wing. Elon Musk being the most obvious example, but also the owner of Nvidia, Jensen Hueng, outwardly praising Trump’s reaction and support of AI growth. These people are not your friends, and they are becoming extremely powerful in an ever growing industry. There appears to be no counter-balance of a left-wing viewpoint coming through from AI companies. The left’s response to the phenomenon has effectively been to demonise it and pretend it isn’t happening.
If you want AI to be developing in a productive direction, someone on the left has to start developing products or regulations that provide an alternative to the right-wing orientated channels. Although there are widespread suspicions AI is a bubble, and it likely is to some degree, it is still being utilized so widely that some AI products will become indispensable. If we start allowing private, predominantly politically ring-wing companies to freely shape our romantic, sexual relationships, I imagine the outcome will not be the utopia we were promised.
At the end of the day this is a money game for those companies, but for the rest of us it is our real lives. As the number of individuals, predominantly male, that use AI chatbots to relieve whatever their internal desires, I wonder what those people will start to behave like in reality too.