I don’t know what I think about AI (“artificial intelligence” for the uninitiated). Like any new advance in technology, especially technology of computers, it has the power to be useful to those of us who use computers daily, or even less often. The search engine I use most often is now advertising that it is “AI powered”. I’m not sure what that means, but it could mean that it searches more efficiently and effectively, generating results faster and more to what the searcher is looking for than the older way. Time will tell. Actually, I’ve never noticed any sort of increase in speed or accuracy of results, but I don’t use search engines that much anyway. If search engines can be powered by AI, other computerized machines may be powered as well, and I’m assuming some already are. Spaceflight could be one application for artificial intelligence: guiding spacecraft, avoiding problems, diagnosing trouble faster than the crew can react, taking steps to minimize downtime, initiating repairs, and so on. I’m certain that if AI had been installed within the HAL 9000 computer in the spaceship Discovery One bound for Jupiter in the movie 2001, A Space Odyssey, it would have detected the anomaly in the AE-35 unit earlier (do you remember that little box they took out of the back of the main communications antenna midway back on the ship?) and HAL wouldn’t have been able to come up with a reason for taking over the ship and killing the crew. [“Open the pod bay doors, HAL.” “I’m sorry, Dave, I can’t do that.”] It might have been interesting to watch AI and HAL fighting (electronically, of course) within the computer mainframe of the ship. Care to place bets on who would win?
But aside from a few humorous observations about AI, the real problem with AI for me is the use of copyrighted material (books, magazines and other forms of written material) as fodder to train AI. I don’t know enough of how AI is trained, or how an AI program is written to understand how the software is “trained,” but I do know I don’t like the idea of using someone else’s material, including mine, to train it, without obtaining permission from the author. A copyright is supposed to provide legal protection for an author from use by someone else. (Illegal uses are things like plagiarism, unauthorized reprinting, unauthorized taking of movie rights to make a movie or TV show, and so on.) Scanning a book or article or story into a computer to “train” it seems like it comes under the heading of ‘unauthorized use’. I believe when I get a copyright for my book, I control its use in all phases. I write books for the simple reason that they should be read and enjoyed (if possible) by others, and not used as food for computers. For this reason, I have inserted the sentence, “This work may not be used in any manner, form, or fashion, in whole or in part, in the development, education, instruction or training of artificial intelligence software without written permission from the author,” in the front of my books after the declaration of copyright. The copyright should protect against this. This declaration assumes that if someone asks me about using my book(s) for AI purposes, that I may think about it and possibly say “yes.” That’s extremely unlikely in any event. AI is so new, I don’t want my books scanned into someone’s computer program to be used for God-knows-what, until effective controls on the process are put into place. Perhaps I’m old-fashioned, but better safe than sorry.
Comments?