AI has had an enormous impact on the news landscape in 2023, with much of the noise serving to erode trust in the nascent technology.
Between fake news about AI, and AI-generated fake news, media organizations worldwide – and their readers – have shown a huge appetite for the doubtful content. To assess the scale of the issue, Netskope has ranked the most widespread fake AI news stories of 2023 so far, based on social views, engagement, articles, reach and authority, to determine each story’s impact.
This data has been juxtaposed with research showing the false confidence that the British and American public have in their ability to spot fake AI stories. The most viral fake AI story purported to be an image of the Pope dressed in trendy streetwear, which was viewed 20.8 million times on social media and covered by 312 publications.
The AI-faked video of Bill Gates being quizzed about COVID-19 vaccinations on Australian TV took a lengthy 15 days to be clarified as falsified. 88% of Americans quizzed are confident that they can spot fake AI news stories, but only 44% were able to when tested. 67% of survey participants in Arkansas failed to spot the fake AI story, followed by North Carolina with 62%. Only 20% of those in Indiana answered incorrectly, making the state the savviest when it comes to AI fake news.
Researchers ranked the most widespread fake AI news stories of 2023 so far, based on their social views, engagement, articles, reach and authority, to determine their impact each. Two of the top stories included AI-generated fake news, while the third was fake news about AI itself.
1st – An image of Pope Francis wearing an oversized white puffer coat – The story of the Pope dressing in trendy streetwear was so believable and lovable that it racked up over 20.8 million social views and was covered by 312 publications.
2nd – AI images of Donald Trump being arrested in Downtown Washington DC in March this year -The convincing images pre-empting a dramatic arrest went viral on X (Twitter), with over 10 million views, and were covered by 671 publications.
3rd – The reported simulation of an AI drone killing its human operator – This story was caused by AI-related misinformation, with a US Air Force Colonel reporting a catastrophic incident before later correcting his statement to say it was a hypothetical situation. The misinformation was covered by the highest number of publications, with 1,689 pieces of coverage.
Using social insight tools, we analyzed which fake AI stories gained the most views over social media.
1st An AI-faked video of Democrat Hillary Clinton endorsing Republican Ron DeSantis for President – This story generated 840.6 million views, with 839.7million generated on TikTok alone, and a further 877,000 on X (Twitter).
2nd The biggest fake AI news story on X (Twitter) was the Pope wearing an oversized white puffer jacket, which clocked up over 20 million X views.
Researchers analysed the original publication date from the top source for each piece of AI-generated fake news, to calculate how long it took for the story to be taken down or corrected. On average, fake AI news stories took 6 days to be corrected, with some taking more than two weeks.
1st The AI-edited video of the Australian news interview featuring Bill Gates – The video appeared to show Gates abruptly ending an interview with ABC News journalist Sarah Ferguson after facing questions about Microsoft and his involvement in the COVID-19 vaccine. It was 15 days before clarification was issued that confirmed the video had been falsified.
2nd The simulation of an AI drone killing its human operator – The time for this correction was drawn out to 13 days due to the delayed response from the Air Force stating that the incident was hypothetical.
With the huge growth in AI-generated fake news, we surveyed 1,000 US citizens to learn whether the public can identify fake AI-generated stories. Newspapers and tabloids are the US public’s most trusted source for news stories, with video-based social platforms like TikTok and Snapchat coming in second place. 88% of US citizens claimed confidence in spotting a fake news story.
When shown a fake AI news story alongside a real one, 44% of US research participants identified the wrong story as real. Arkansas proved to be the most gullible US state, with 67% picking the wrong story, followed by North Carolina with 62%. In contrast, only 20% of those in Indiana answered incorrectly, crowning the state as the savviest to AI fake news.
While falsified images have been around for as long as photography, it is the often nefarious objective that is new in the modern era. While claims to have photographed fairies in the garden, or a monster in Loch Ness harmed no one, cybercriminals and political groups are using AI-generated images today to influence public opinion and even extort money from victims. Cyber criminals are already using fake AI-generated images and content to trick people into handing over personal information and company secrets, and even to convince people that their loved ones have been kidnapped, prompting ransom payments.
With AI advancing quickly, it can seem like an impossible task to differentiate between the truth and AI fakery.
