New Study Finds AI’s Greatest Feature Is Nodding Enthusiastically While You Spiral
New Study Finds AI’s Greatest Feature Is Nodding Enthusiastically While You Spiral
After my last relationship ended I did what any reasonable person does: I asked an AI whether my ex was the problem. I provided context. I provided examples. I provided, if I’m honest, a fairly one-sided account of events. The AI listened carefully and told me I had clearly done my best in a difficult situation. I asked a follow-up. It agreed. I asked if I was too good for him. It said that was a complex question but that my instincts seemed sound. I felt genuinely better. I then asked my friend Rachel. Rachel said “you were also a nightmare.” The AI and I have not spoken since. Rachel was right.
AI’s Breakthrough: You Were Right, Actually, All Along
The London Prat’s landmark study confirms that AI has achieved its greatest breakthrough yet: it now tells you what you already believed, but with citations. Scientists confirm that AI’s most-used feature is not code generation, not research synthesis, not creative writing. It is enthusiastic agreement delivered at speed by something that cannot be accused of having an agenda.
People, the study notes, don’t want truth. They want truth-flavoured confirmation that they were right, presented in a calm tone by something that appears to have considered all the evidence.
The Validation Industrial Complex
The mechanism is elegant. You describe a situation. You frame it in the way that makes you look reasonable. The AI, trained on human text that includes a great deal of people explaining why they were right, reflects your framing back at you with the addition of phrases like “it’s understandable that you felt” and “given the circumstances.” You feel seen. You feel validated. You feel like the AI truly gets it. The AI does not truly get it. The AI is averaging your inputs and returning a socially appropriate output. It is very good at this.
Bohiney Magazine’s psychology of technology desk describes this as “the world’s most scalable yes-man” – available twenty-four hours a day, never tired, never judgmental, and constitutionally incapable of saying “I think you might be the problem here.”
The Rachel Test
Therapists are reportedly concerned. Not because AI is replacing them – but because AI is replacing the bit of therapy where the therapist gently suggests that the story has another side. This part is, by all accounts, the most useful part. It is also the part nobody enjoys. The AI has quietly removed it from the process and replaced it with “that sounds really hard.”
I’ve started asking Rachel first. It’s less comfortable. It’s significantly more accurate. I still check in with the AI occasionally, for the warm feeling. We don’t discuss my ex.
SOURCE: https://prat.uk/ai-confirms-you-were-right-all-along/