There are multiple inaccuracies. In particular, the article implies that tokens are universal to all AI—which is incorrect. For example, the statement that "AI systems depend on visual tokens to make sense of images" is wrong. This might be true of vision transformers, but not of CNNs.

Tokens are the very non-secret building blocks of any "language". They themselves are not the "language" of AI any more than the embeddings are. Though at least you could reduce it down to the embeddings for the purpose of simplicity.