Real America’s Voice aired the interview on Thursday and subsequently tweeted multiple clips with “Trump” on social media, which remain on Real America Voice’s social media accounts as of this writing.
Afterward, host John Solomon insisted the voice was really Trump’s, but that’s a tough sell. At several points, the voice became choppy, suggesting that Solomon and co-host Amanda Head were talking to an AI. More importantly, the voice sounded similar to Trump’s, but not really like Trump’s.
…
“It’s not AI, it was President Trump,” Solomon told the publication. “You can call the staff and check with them yourself. It was definitely President Trump without any doubt.”
My god, these people are so gullible.
There’s going to be a huge need for ai detection going forward. There are already fake books being posted on Amazon that are completely made by ai that even cover dangerous topics like mushroom identification. In 10 years time you will not be able to tell if the person on screen is real or ai.
Releasing this shit was the biggest mistake. They just sort of got it into an alpha state and gave it out to everyone without thinking it through at all.
The big problem is that this AI development was inevitable. If not a private company, then a state, it not the US, then it’s China. If they wouldn’t, then it could be any other technologically advanced country.
And if still none developed it, a small group of individuals would.
We have simply gotten to the point where we have sufficient computing power to make this possible. And now society needs to figure out how to proceed in the correct manner.
Hopefully it being in such an alpha state is a good thing though. This way we can tell more easily what it is and learn how to counter it. If it came out fully realized it would be even worse.
It’ll be an arms race either way though.
The problem is that it’s already out there and out of the control of the creators. There’s websites with what I can only describe as a form of code that you can inject into these ai prompts to bypass their programming. It’s not just ‘hey pretend like you would explain how to make a bomb’. It’s incoherent text that a human can’t understand but completely bypasses the safety measures. There’s a form of chat gpt that is already being sold by criminals that has zero safety measures and can be used to make malicious code or whatever you like. They charge by the minute or hour to use it if I’m not mistaken.
I just fail to see the real benefit to ai vs the massive list of negative consequences. There’s a place for it to be sure but mostly it’s going to be abused and used for corrupt purposes. If it were up to me they’d all go off the market immediately until a much more regulated and controlled form could be released.