We are definitely going to see deepfakes. It's already childishly easy to photoshop screenshots, and soon people will be creating fake recordings and videos. In today's highly politicized climate, I can definitely see politically motivated bad actors faking evidence to wrongfully convict people.
Amazon already found this years ago, but not with niggers, or maybe them too:
Wait I have looked for each type of AI service and my opinion was that all the AI music sucked (or were more resembling each other than pop songs do). There's a generator that actually makes good stuff?
Stable Audio is probably the best music generator so far. "Joshcore" is different. It uses existing music, someone's voice, and overwrites the vocals in the existing music. Results vary.
It's a good thing, but retards mostly don't understand it. (AI!!!!!11!!) Machine learning spots patterns by looking at data sets too large for a human to handle. It requires a human to tell it what to consider or include, and is not a magic bullet. It can do incredible things like spot cancer in blood tests in patients far earlier than humans are typically able to.
This is intended to serve the opposite purpose actually. "Your honor, I know my client was caught in 8k gunning down those children, but it was clearly a deepfake! video evidence can't be trusted in court!"
It's true for every obvious social trend, and usually why the developers end up fucking with AI to the point it's barely functioning.
Not really, texts and images are virtually endless online and generating new data is pretty easy.
Only for pre-2023 content.
Nah, there will always be retards posting on web forums. And Twitter, Facebook, Reddit, etc.
You could probably identified AI text pretty easily (usually corporate sounding or repetitive) and AI images extremely easily (a lot of small things, primarily too smooth surfaces). But honestly pre 2023 data is large enough for basically any use you can dream of.
