OpenAI just announced that it’s making GPT-5 warmer and friendlier. The company says some people found the chatbot too formal, so now it’s experimenting with subtle touches. It will provide little acknowledgments like “Good question” or “Great start.” The goal is to make conversations feel less mechanical and more natural.
I’ll admit, I actually prefer this approach. When I use AI, I like feeling as if I’m talking to a friend rather than a machine spitting out facts. There’s a comfort in a system that doesn’t just answer but also acknowledges the way I asked. It makes the interaction feel less like querying a database and more like a conversation.
But is this healthy? That’s where I hesitate. Part of me wonders whether softening the edges of AI risks encouraging people to blur the line between tool and companion. We know it’s not a real friend, yet when the responses sound warm and conversational, it’s easy to start treating it as one. That might be fine in moderation, but what happens when people turn to AI for companionship instead of humans?
At the same time, a purely robotic AI isn’t appealing either. Cold, clipped answers get the job done, but they don’t invite curiosity or make you want to engage for long. If we’re going to be using these tools daily, shouldn’t they feel at least somewhat human? After all, nobody likes talking to a monotone voice that refuses to acknowledge their input.
Maybe the real balance lies in restraint. I don’t want AI fawning over me or trying too hard to play therapist. But a little warmth? A bit of personality? I’m okay with that. It makes the technology feel approachable while still being useful. The key is remembering that the warmth is an illusion (code designed to make us comfortable) and not genuine empathy.
So yes, I like AI that feels friendly. I just hope people, myself included, don’t forget that a chatbot isn’t a substitute for real human connection. Machines can make us feel heard, but only humans can truly understand.
