Deepfakes might be busted as they’re going to usually carry these 4 flaws, professional says

Convincing deepfakes threaten to interfere with the UK general election and spark political chaos.

As artificial intelligence (AI) continually improves, it’s getting harder and harder to distinguish between what’s real and what’s not. Already, we have seen computer-generated audio clips of Sir Keir Starmer supposedly swearing at staff, and a viral soundbite of London major Sadiq Khan making controversial remarks about Remembrance weekend.

While both clips were outed as deepfakes, it’s clear AI has the potential to sway an election through fake news. Luckily, founder of 10 Yetis Digital Andy Barr, says AI-generated content can be busted as it likely has four hidden flaws.

READ MORE: AI chatbots found to allow and encourage Nazi, sexist and racist views

Stay right here on the Daily Star for the latest UK news



Labour Party leader Sir Keir Starmer has been victim to AI deepfakes
(Image: PA)

Reputable source

“My own tween kids quite often share funny videos that are clearly fake and AI-generated of celebrities, and I tell them to check the source of the account posting the video,” he said. “Does it have a verified tick? Does it appear as a news item from a reputable source such as The BBC or a similarly strong global media outlet?”

The expert also urged social media users to think about the ‘content’ they’re viewing. If you’re watching a video where Rishi Sunak is rapping his manifesto to Nicki Minak’s Superbass – it might be worth double-checking its credibility.



AI-generated content tends to be posted on suspicious social media accounts (Stock photo)
(Image: Getty Images/iStockphoto)

Realistic content

Barr cites a real-life example of this, after an AI-generated video of former POTUS Barack Obama using ‘very strong language and swear words’ to describe Donald Trump went viral. “It’s clearly very unlike his usual tone and style and is obviously fake,” he added.



Deepfakes can track facial movements from a simple picture – making them ultra realistic (Stock photo)
(Image: Getty Images)

Account following

The tech ace says it’s worth looking at the account posting the content. If they’re mainly followed by ‘faceless accounts’ or people ‘claiming to be crypto specialists’ it’s probably so a trustworthy source.



Sadiq Khan says he’s ‘concerned’ about AI deepfakes – after being the target of one himself
(Image: Getty Images)

Too good to be true

“You also need to apply the old adage of: ‘does it seem too good to be true?’,” Barr concluded. “If so, it probably is fake.”

According to Sky News, a UK government spokesperson addressed the looming concern of AI, stating: “Security is paramount and we are well-prepared to ensure the integrity of the election with robust systems in place to protect against any potential interference. The National Security Act contains tools to tackle deepfake election threats and social media platforms should also proactively take action against state-sponsored content aimed at interfering with the election.”

For more incredible stories from the Daily Star, make sure you sign up to one of our newsletters

artificial intelligenceBBCDonald TrumpLondon