AI OMG

The other day I watched George W. Bush, Bill Clinton, Barack Obama, and Joe Biden eat tacos and laugh. Mocking, of course, the Cheeto in Chief’s latest nickname: TACO. (Trump Always Chickens Out.)

Then a brief speech by Denzel Washington took over my screen. After that, Kelly Clarkson tried to sell me weight loss gummies.

The catch: Not one of these videos was real. They were all AI (Artificial Intelligence) generated and in Clarkson’s case, extremely damaging to her reputation. In fact, she, Melissa McCarthy and others who ‘star’ in fake testimonials can’t get them taken down.

It’s a whole new world out there, trying to figure out what’s real and what isn’t.

Check Your Sources

We journalists (and former journalists) have always double and triple checked everything to make sure it was true. It’s part of the job. But millions of people don’t know or don’t care to confirm. Especially if it’s about something they agree with. Why check the truth of what you believe? Actually, that’s when you should be even more vigilant.

Last week, several of my Facebook friends shared a post about tick removal. “A nurse discovered”, it began. A cotton ball soaked in Dawn dish detergent removes a tick. No, it does not. No one with an ounce of knowledge recommends this but it has gone viral again. It comes around every year or so. And it’s not correct. In fact, a tick might burrow deeper if you try it.

So, how do you check for validity? First, look at the website. Is it NBC or NBCrealnews, for example. Fakers will often make an URL that’s close to a real thing, but not quite the same. Is at a site you’ve even heard of? Or is it a YouTube video made by someone random or an ‘influencer.’ Are there credible sources named or just “a scientist said?” Real people stand behind their claims with their names and facts. I recently saw a viral video of a woman who claimed carrots aren’t real food. A stranger ‘discovering’ the opposite of an accepted fact is usually an idiot hoping for attention. Look for confirmation by a trusted source.

I often go to Snopes.com. They’re independent fact-checkers and they look into the wildest, weirdest stuff as well as everyday false news. They dig deep until they find out whether it’s authentic or made up. The Associated Press also has a fact checker, as does CNN. Rachel Gilmore is a Canadian fact checker, mostly covering our country’s political stuff and unraveling any BS behind it.

AI VO WTF?

The other night, a young salesperson was in our house, saving us money. For real! He asked what we do – voice acting – and he played us an app he thought might be AI.

AI voices used to be easily picked out because the “voice” never took a breath. That’s not the case anymore. There are AI voice programs that include breaths to make them sound more authentic. We can still tell when the voice isn’t human because the emphasis is often wrong. This guy’s app definitely used AI. The speech followed a repeated pattern and came out too even and perfect. He hated it and wished it was real.

In fact, recent studies show mistakes, especially spelling errors in the written word, signal it was made by a human. College students are making reports with AI and then sprinkling in typos. Some companies are apparently building spelling mistakes into their ChatBots to make them appear human., too

The Upside

I’ve replaced Google searches with ChatGPT. The AI program brings a broader range of answers without needing a million clicks on websites. Still, you can’t trust it 100% and skepticism must be applied.

AI has its uses. Just not in the creative space. Morning Brew reports many companies that jumped on the bandwagon are now rethinking AI. They didn’t want to seem left behind but now they’re experiencing issues. KPMG, for example, found many employees were uploading sensitive client information and making it public. Oops. Some firms that cleaned out HR departments for AI are reversing course. After all, HR stands for “Human” Resources.

And to prove that making an artificial intelligence is difficult, take the case of Natasha. This AI building company attracted investment from Microsoft. But instead of using AI it was relying on the work of 700 human engineers from India. They lasted for eight years before being exposed this month. The company has filed for bankruptcy. Imagine pretending to be fake by using real people!

I had the opportunity to talk to young editors at the Western Gazette – Western University – last week. They’re determined to search for truth in the world of news and information. They gave me hope that real journalists will continue to exist along with original writing and reporting. Fact-checking and verifying sources isn’t dead. It’s just that the Internet has turned everyone into a journalist. And most people aren’t good at it.

The Clone Problem

A good friend is in a fight over the unauthorized use of her voice for AI. She happened to hear herself doing a voice-over that she had never recorded. Her contract with the company clearly stated that she didn’t give permission for her voice to be used this way. Still, she’s had to hire a lawyer and it’s just dragging on.

We’re keeping our fingers crossed that our friend will be properly compensated for this theft. Meantime, AI still hasn’t quite caught up to humans. People like to listen to people. You can purchase an audiobook and never know it’s voiced by AI until you start to listen to it. That isn’t right. And it’s just one of many battles left to be won. Real voices, faces, and fingertips must continue to fight the artificially intelligent ones.