x
Breaking News
More () »

How to know what's real and fake with AI

Most AI technology isn’t foolproof yet, but spotting deepfakes and other AI-generated content requires attention to detail and critical thinking.

Chris Hrapsky

Play Video

Close Video

Published: 10:24 PM CDT May 8, 2024
Updated: 5:33 AM CDT May 9, 2024

We've officially entered an age where seeing is no longer believing. Where images, videos and voices can be entirely created by artificial intelligence, and it's starting to show up everywhere.

In New Hampshire, a deepfake robocall with President Joe Biden’s voice urged voters not to vote in the primary election.

A political attack ad in Iowa used artificial intelligence to mimic former President Donald Trump’s voice to narrate a social media post he had written.

In Russia, the real President Vladimir Putin took a question during an interview from an AI-generated version of Putin.

It’s amazing, if not unnerving how far artificial intelligence has come in just the last year. 

Software like Sora and Vidu are not ready for consumer use yet, but presentations show they can turn words into hyper-realistic videos.

I asked Michelle Gilbert, a principal cloud solution architect at Microsoft, if she thinks people should be afraid of AI?

“I don't think so at all,” said Gilbert. “I think it's incredibly powerful. Get comfortable with AI, 'cause it’s here to stay.”

She is an expert on Copilot, Microsoft’s latest AI companion to its Bing search engine. Copilot can create images and even compose a completed song based on the prompt you give it.

Along with content generation, Copilot uses OpenAI's ChatGPT technology and integrates it with other Microsoft products.

Another Microsoft project called VASA-1 can turn a single AI-generated image and turn it into a 3D talking video.

Ethically, Microsoft says it’s against the tech used for deepfakes or deception, and researchers say this technology might be used in the future to spot forgeries.

Before You Leave, Check This Out