FAIR is not a standard, although the acronym is frequently used in that context. The GO FAIR view is that standards are needed for the Internet of FAIR Data and Services and that ideally, standards, API’s and protocols are developed ‘following FAIR guiding principles’.
FAIR is not equivalent to open (and open is not equivalent to ‘free’): There are many reasons why data may be non-open and only available under certain conditions to certain users, including machines. As long as the accessibility conditions are properly described, non-open data can be entirely FAIR. Reciprocally, fully open and unrestricted data may score very low in FAIR metrics as they may for instance be non-actionable for machines.
FAIR principles do not, in themselves, cover the crucial aspects of intrinsic data quality or ethics. However, FAIR guiding principles request that optimal care is taken to enable users to determine the ‘usefulness’ (for their purpose) of the data and other research objects they find, which includes rich, machine readable provenance. Obviously, user defined metadata and comments on existing research objects will be increasingly useful to judge the reusability of the research objects.