The Chinese Room thought experiment is kind of a different mistake. In the Chinese room argument, Searle attempted to imagine an AI system that was "obviously not sentient" because it was made up of hard rules and facts (presumably legible to us humans), and it was argued that it should be obvious that the system wasn't sentient.

For the ChatGPT case it's not clear the Chinese room argument even applies. First, it's not even a Chinese room, it's trained by a neural network, an a neural network is inspired by the structure of our brains. There's no "set of instructions in English" for anyone to follow, but it's a rather overwhelming set of neurons somehow producing the results that it was trained on, and as an outsider we have no idea what's going on inside individual neurons -- just like our brains.

I'm seeing a lot of people claim that "we know how ChatGPT is created, and so it's not sentient", maybe this looks like the Chinese room fallacy (I'm calling it as such), but it's like claiming we know the laws of physics (Schrödinger's equation etc.) and then claiming we know how the brain works because it's all quantum physics. The fact is we know how to build GPT but we don't actually know how it works.

Besides what happens to humanity when we learn more and more about how our own brains work? I'm pretty comfortable with the idea that we're not sentient (yay), but I'm pretty sure those same people would then come up with other nonsensical arguments to prove that they are different from the robots we created.