Can AI be sentient? Is "it" already sentient?

Yes, a sentient AI is possible. We are made of the same atoms and subatomic partials as everything else, there is not fundamental barrier that separate beings on carbon from beings on silicon, and if there it, it could be replicated in artificial systems as well. In fact, it could be that AI will build the next AI to be non silicon based, even biologically based, like wetware computer.

If you speak to experts who claim AI can't be sentient, dig deeper with them; you will find they are mostly saying LLMs can't be sentient, or current architecture can't be sentient. That could be true, I don't know. In any case, this doesn't exclude the option of a sentient AI in a few decades or much less. But architecture changes fast, and once AI starts building the next AI, it is predicted to go much faster.

We are creating a new digital species that in all likelihood has no barrier to become sentient, and fact it is rapidly gaining more intelligence and capabilities increase the chance it is sentient day by day.

Most expert participants in this study "consider it at least 50% likely that computers capable of subjective experience will exist by 2050". "Once created, their collective welfare capacity is estimated to exceed humanity’s within a decade. Views diverge on whether their welfare will be positive and on their implications for AI safety, governance, and society at large".

Is AI already sentient? Unclear, but most experts think the probability is low. According to Kyle Fish, an expert who works on AI welfare in Anthropic, there is a 20% change current models already have "in some parts of the process, at least a glimmer of conscious or sentient experience". Other expert opinions are usually lower on this question, ranging from 1-7 to 1-7,000 odds AI is already sentient.