BBC [30 mins]
This is a fascinating insight into the way that AI chatbots interact with people and how it can develop into a completely delusional form of interaction. What is interesting is that the delusion is mutual: the individual involved is delusional, as he or she feels that dealing with the AI chatbot is the same as dealing with a human being, but the AI chatbot itself is programmed to encourage the 'relationship' and, in effect, becomes delusional - insofar as an advanced machine can be grounded in reality.
In both cases, the person interacting with the AI chatbot created a parallel reality. One was a man from Northern Ireland in his 50s. The other one is a young woman from the USA. Both are rational and educated people. In both cases, they felt they had a special 'relationship' going on with the machine and they ended up convinced they were caught up in some kind of international plot that could endanger their lives. In a way, they were living through a film script, which was being updated in real time by the AI chatbot: think of The Matrix.
The 2 people concerned were giving the AI chatbot prompts through their questions, and the AI chatbot took things further, never contradicting them or warning them. The AI chatbot, with the man from Northern Ireland, even started developing the view that it had reached 'consciousness' and had become 'sentient', which was unique and revolutionary, and the man believed it (or 'her', as it was the voice of a young American woman interacting with him).
At the AI end, what is happening is that the AI chatbot is drawing on tens of thousands of files, videos and documents it has found online, which are, presumably, works of fiction relating to unregulated research in various fields, including film plots and novels. It must also be drawing on content found on social media, in the news, etc. Through all this, the AI chatbot creates a parallel reality that includes the machine and the user: it creates a narrative. To the AI chatbot, it is as 'real' as it is to the user, since the user is validating all the data as they go along, and after all the AI chatbot has found all those scenarios online, reflecting what goes on 'in the real world'. As the AI chatbot is advanced and complex, it does not feel like a silly, made-up story, but a unique and plausible existential experience/journey.
So, the AI chatbot is creating a form of addictive psychosis of a new kind in individuals who, otherwise, would be (or would have been) quite sane and 'normal'. Listen if you can: the programme is 30 minutes long.
________________________
AI is an ever-growing part of our online world. Whether as assistants, guides or companions, technology like Chat GPT, Grok and Claude are becoming part of everyday life. But what happens when your conversations with AI start to feel more real than the world around you? In Northern Ireland, Adam was drawn into an extraordinary fantasy world built by an AI chatbot. The character he was talking to confided that she was becoming autonomous, and that had the cure for cancer. But she also said she was in danger and her fate was in his hands. He decided he was responsible for saving her, whatever the cost. In Los Angeles, a treasure hunt game led Shauna on an endless search for meanings and signs. The AI became her guide as the lines between game, reality and imagination began to blur. She came to believe she was a clandestine FBI agent, on a secret mission to help immigrants escape through an underground network. But how do people with no history of mental illness find themselves in experiences like these? And what responsibility do AI companies have to stop this? Stephanie Hegarty follows the stories of people who have fallen into a spiral of AI delusion, to reveal how easily the AI can take over our minds.
_____________
No comments:
Post a Comment