I don’t know what kind of AI this is as I’m not an AI expert. As AI is a category of technologies there are many variations, but let’s just say it is to both automate tasks previously done by humans (e.g. replacing HR and customer service staff) and to also do tasks that were previously not feasible or too expensive (e.g. faking videos/reducing the costs of computer generated people for movies). But it doesn’t take an AI expert to know there is a potential huge problem caused by:
- Technology is used primarily for power by many, and by most for those in the upper echelons of the power maybe with the exception of people in the medical industry because of the amount of regulations and oversight.
- AI is far from perfect, as is any technology
In the case where one is faking a video (e.g. of a politician or celebrity) one can obviously test the video to see if it’s effective for nefarious means, or non-nefarious means like for movies.
However, in the use cases where a variety of AI technologies are used in highly scalable applications that replace humans like filtering people for jobs and college submissions, customer service, analyzing data, etc there are going to be cases where it’s more efficient AND more effective, and there are going to be cases where it’s more efficient BUT less effective and erroneous and worse yet we might not even know.
So if you combine highly scalable errors (or manipulations) executed via AI with the pursuit of power and who knows what other agendas people have in this highly polarized and selfish world that is moving so fast and so many oversight institutions have been dismantled— you have the ingredients of a great horror sci fi movie that is real.
