There are legitimate concerns about the rapid rise of innovative technologies like artificial intelligence and how false or manipulated images or audio created by AI can be exploited to deceive the general public.

One recent example making the rounds on social media is a purported photo of former first lady Michelle Obama wearing an anti-Trump T-shirt, which is a digitally altered image derived from a photo of her wearing a shirt promoting her newest book, Business Times reported.

The fabricated image, which went viral and did indeed trick many users into believing that it was real, highlights the need for everybody to be vigilant and circumspect about the images and audio they see and hear online in the emerging AI era.

Viral fake image of Michelle Obama exposed

Reuters reported on the viral image of the former first lady Obama appearing to wear an anti-Trump shirt that was blue and featured the words "NOPE, Not Again," with the "O" in "NOPE" stylized with blonde hair and a red tie as a caricature of former President Donald Trump.

Who agrees with @MichelleObama? 💙 pic.twitter.com/3CrFZyDWqZ

— Jon Cooper (@joncoopertweets) November 11, 2023

That image is not real, however, but rather is a digitally altered version of a real photo that Obama previously posted to social media to promote her latest book, "The Light We Carry."

In the original photo posted in March, the former first lady wore a yellow shirt that featured the initials of the book title, "TLWC," and was intended as promotional merchandise that supporters could purchase and wear.

I’m excited to share The Light We Carry merch with all of you! Check out the https://t.co/zDXupr2f6K for everything from shirts, hoodies, beanies, and more. Snap a picture of what you get! Can’t wait to see you rocking it. ✨ pic.twitter.com/IDS2jbyg3c

— Michelle Obama (@MichelleObama) March 14, 2023

Already being used to deceive voters

According to WRAL, the increased usage of AI-generated or manipulated imagery and audio and the deceptive impact those altered images and audio can have on the general public has sparked real worries among some, particularly in the political realm.

There have already been several instances of AI-generated or manipulated attack ads against political candidates, whether created by a campaign or some outside or affiliated group, that have purported to show a candidate saying or doing things that they never actually said or did in reality.

"Being an informed electorate just got a lot harder," Amanda Sturgill, a journalism professor at Elon University, said of the use of AI in political ads. "You can actually put stuff into videos or into photos that wasn’t there to begin with. But you can also make something whole cloth out of nothing."

To regulate AI usage or not?

A debate has now emerged about what, if anything, the government should do to try to regulate the use of AI in political campaigns or to target public figures with dishonest imagery and audio, which have also been dubbed as "deepfakes."

Some believe that there should be laws prohibiting the use of AI, or at least requiring public disclosure of its use, while others counter that any such regulation would violate the First Amendment's protections for free speech.

As things stand now, people can do little more than remain vigilant and look for telltale signs of AI generation or manipulation, such as abnormally shaped appendages, especially fingers, and unnatural or exaggerated facial expressions and movements, as well as mistakes in the background like nonsensical letters in place of actual words on signs.