OpenAI's high-minded approach to AI-human relationships ignores reality

OpenAI’s Head of Model and Behavior Policy, Joanne Jang, has penned a blog post on X about human-AI relationships, offering some well-considered ideas about the subject and how OpenAI approaches the issues surrounding it. Essentially, as AI models get better at imitating life and engaging in conversation, people are starting to treat AI chatbots like they are also people. It makes sense that OpenAI would want to make it clear that they are aware of it and are incorporating the facts into their plans.

But the thoughtful, nuanced approach, including designing models that feel helpful and kind, but not sentient, misses something crucial. No matter how clear-eyed and careful Jang tries to be, people having emotional connections with AI, an occasional outlier event, or a future hypothetical, it's happening now, and it seems to be happening quite a lot.

OpenAI may have been caught ...