I cannot tell if this is misguided or evil:
AIs are no longer mere tools like smartphones and electric cars, and we cannot treat them in the same way as mindless technologies. A new dawn is breaking.
We need an AI rights movement | The Hill
It might just be clueless because the author seems to suggest that because we cannot always tell why the results of a specific deep learning model occur then we cannot know if real thought has happened. That is nonsense, at least as it is applied to the "AI" tools we have today. We know precisely how they work -- they are simply predicting what come next in a string of text or in a visual medium. The rules are ones we provided and we know that there is no thought because we know that they cannot produce anything other than regurgitations of what exists in their training data sets. Heck, even the best of the programs are often caught plagiarizing or copying material wholesale from their training sets. Pretending that these autocompletes on steroids could approach sentience and deserve rights is, to put it mildly, putting the cart before the horse. In fact, the cart is a grove of trees and a pile of unprocessed iron. The horse is a gleam in mare's eye.
But this could be something more sinister. Some of the people quoted in the piece appear to have financial stakes in the success of the current crop of AI generative tools. But those tools today depend upon taking material from artists and writers and reusing it without payment or permission. The ethics of that is dubious at best, and the legality of the process is being contested in court as we speak. However, if chatbots had rights, if their owners (because you can be certain that none of the rights would be the right to not be owned and used by their programmers) could even convince people that they were closer to human beings than autocomplete?
Well. Suddenly attempts to force these companies to behave ethically with respect to artists or to regulate them so as to ease the disruption to society? Those aren't issues for democratic debates and the court of law. No, those are violations of rights at worst, unjust tampering with a vibrant new class of potentiality that we should leave to our technological betters. Betters who will surely use their newfound wealth to shepherd in a glorious AI future. For themselves, at least.
One day, we may have to deal with these issues (though I am skeptical. If you look at the history of computer science, we have been ten to twenty years from true artificial intelligence since before I was born, and I learned to program in FORTRAN. I'm old, is the point.), but chat bots and image generation software come nowhere near requiring such a rethink. They are just advanced version of common tools, tools that we understand perfectly well how they operate. And who they benefit. And who pretending the bots are something other than tools will benefit even more.
Want more oddities like this? You can follow my RSS Feed or newsletter.