3.1: Dataset: Add a "language" field

[bact]

AI team meeting 2024-04-24 made a decision to explore about the addition of a field/fields about language/linguistic properties of a dataset and a model.

While some repository may made this language filed a mandatory metadata,
we should keep in mind that not every datasets are having language aspect (or even human-associated data)
and not every models are language models.

• No Assertion or Not Relevant could be a possible value.
• IETF BCP 47 language tag can be considered as a standardized code

More discussion to follow.