AI scams: Only 1.5% of Brits can identify all the real vs deepfake celebrity videos in this quiz

Artificial Intelligence
Share
The majority of people thought the deep fake of Boris Johnson was real

When shown a series of 8 short video clips featuring both real and deepfake celebrities, only 1.5% of Brits were able to correctly identify them all.

This latest research from personal finance comparison site finder.com highlights the danger from the rise in deepfake scams in the UK. Participants were shown a series of 5-second video clips of influential people including Boris Johnson, Mark Zuckerberg, Barack Obama and Kim Kardashian. 

The results were startling, with over 98% of 2,000 participants unable to correctly identify all 8 video clips. Just 56% of participants were able to correctly identify more than half of the video clips, and less than 10% were able to score more than 6 out of 8. 

The 2 most convincing fake video clips were of Boris Johnson and Jeremy Corbyn, with 63% and 61% of respondents respectively believing that the deepfakes were real. This is particularly concerning due to the level of influence these individuals hold, and the fact that they are currently 2 of the most famous and recognisable politicians in the UK.

The deepfakes were so convincing that many Brits thought the real videos were fake. 30% and 28% of people mistakenly thought that real videos of Donald Trump and Barack Obama were not real.

Older generations scored higher than their younger counterparts

Perhaps surprisingly, the younger age groups didn’t guess as many correctly as their older counterparts. Those in gen Z (aged 18-23) and millennials (aged 24-42) were left with an average score of 4.2 and 4.5 respectively. 

On the other hand, those in gen X (aged 43-54) scored an average of 4.7 out of 8, and baby boomers scored an average of 4.8. The silent generation (aged 74+) were also far ahead of those in gen Z, with an average score of 4.7. 

Commenting on the findings, Liz Edwards, editor-in-chief at finder.com, says:

“The proportion of deepfakes among all fraud cases has grown considerably in the UK over the last year, rising from 1.2% of all fraud cases in 2022 to 5.9% in Q1 of 2023, according to a recent report. The more time goes on and this technology develops, the more realistic these deepfakes become, and this poses a real threat to consumers who may not even be aware that this technology exists. 

“As our new research shows, most Brits are unable to consistently tell the difference between real and AI-generated videos. If the technology is used to manipulate messages from influential individuals –  to show them asking for money or promoting a scam, for example – this is where the real danger comes in.”

See how many you can guess correctly here: https://www.finder.com/uk/can-you-identify-the-deepfake-videos 

3 ways to prevent being scammed by deepfakes:

  1. Do your own independent research.  If you encounter any form of content which is encouraging you to part with your money, make sure you conduct your own investigation into the opportunity. A quick Google search could reveal whether the video is real or not. You should also check with a respected news source like BBC News to see whether there are reports backing up what the person has said, or reports of an AI scam involving them. Taking a few minutes to do some research could help to avoid losing your money.

  1. Study the video in detail. There are a few key details to look out for in deepfake videos. For example, is their mouth moving completely in sync with their words? Does their voice match the exact tone and dialect of the actual person? You can also check for features such as the shadow movement of the person in the video to ensure that it matches. If anything looks suspicious, it’s best to assume it’s not real. 

  1. Listen carefully to what they’re saying – Although these deepfake videos look realistic, a lot of the time the content will have been written by AI technology, so listen carefully and try to decipher whether this really sounds like something a human would say. 

Chris Price
For latest tech stories go to TechDigest.tv