Academia is full of things you "should" like data sharing, preregistrations, and all kinds of statements about honesty. However, there is almost no incentive provided for such actions and I wonder how effective the measures are in the first place. It’s like running a company by metrics only. It’s great for management but stifles innovation. At the same time, more rules do not stop truly malicious actors, they will find a workaround anyway.
Many scientists think of their data as proprietary, and want to prevent other researchers from getting it and publishing on it. This is at odds with any lip service they may pay to open science, and, often, at odds with the explicit requirements of their funders. As noted in this article, they often deny reasonable requests for data, or just ignore the requests altogether. Of course, everybody wants access to other people's data.