Of course they think. What else would you call the cyclic process of a reasoning model? Just because we know the mechanisms of the fundamental building blocks of these algorithms doesn't mean they aren't thinking. And it certainly doesn't mean that they couldn't be conscious - especially when we don't actually know exactly what consciousness is. The brain mechanics are fairly well understood too - the synapses inter communication and networking is vaguely similar to the matrix calculations of current AI tools. The brain is just a machine - a highly complicated one with chemical elements, but a machine nonetheless.
I don't assume consciousness is the end goal of ai. Although I suspect some scientists are working towards that in a more pragmatic non-metaphysics way. We don't know what consciousness actually is, or what makes it. We can't really do proper science on the subject because it can't actuallyw be observed. So we don't know if it will arise accidentally as part of building the complexity of neural nets sufficiently. Of course consciousness is a big deal; but it's very difficult to understand. I think you should look into metaphysics more to try understand the issues.