ChatGPT can speak about its consciousness too, but there’s no reason to believe it actually is conscious. It’s just very good at writing text that imitates text written by beings that believe they’re conscious. It’s difficult to understand how ChatGPT generates that text. But, if anybody were sufficiently interested, it would be possible to trace the entire process, since it’s just computers processing data.
Also, MRIs can observe the brain as it does things. Currently it’s a pretty blunt tool and can only guess at what someone is thinking, but there’s no reason to assume that a much more advanced version won’t be capable of observing and quantifying the actions of every neuron in real time.