What we are not being told about AI
There are probably thousands of things we’re not being told.
But this is what interests me at the moment:
AI’s access to copyrighted information.
It’s already an issue. If AI accesses a published book, for instance, and drops the text into its database, is that stealing? Does the owner of the AI owe royalties to the author?
Should the AI have asked for permission to access and take the text up front?
And even deeper—how did AI access that copyrighted book in the first place? I would say that’s a big enchilada.
How does AI crawl into a book? It’s certainly not buying it. Should we assume there is some giant database that contains the text of “all books?”
And if AI can’t get access to copyrighted books, how can it possibly answer millions of possible questions with a reliable degree of accuracy and completeness?
Going still deeper, are there now AIs that have access to classified documents possessed by government agencies? CIA? NSA? MI-6? If so, these AIs would obviously be off-limits to almost everyone.
If there are such AIs, then we’re not only talking about government secrecy, we’re also talking about secret AIs that contain these secrets. “Meta-secrets.”
Are you an author? Have you been publishing on the Internet? If so, do AIs have the text of everything you’ve written? If so, what is it doing with that information? How is it integrating it into other information it possesses on the subjects you write about? Does it consider your work “outlier?” Of no interest? Of minor interest? Of special interest?
Do AIs keep lists of authors to whom they assign various rankings? According to what criteria?