I would say, treat the AI like a human viewing said data/material, but unlike the majority of humans, most dont have Kim Peek levels of observation and recall,

I would say, treat the AI like a human viewing said data/material, but unlike the majority of humans, most dont have Kim Peek levels of observation and recall, so should there be some sort of expiration of data built into AI's and if so, should it be a blanket cut off date for all data an AI has been exposed to, or allow for some specialisation like a human might have in order to fulfil their occupation? I'm also aware that search engines have access to data, most humans do not, which gives the search engines an advantage, think accessing ft.com for example, so should AI's have access to data in the context of a search engine or as a human? Its tough, I want to give AI's full access to see what the tech is truly capable of, but I'm aware this will lead to more global monopolies over and above the existing search engine monopolies allowed to dominate in a country or territory, which will harm smaller AI entities.