A fake, apparently AI-generated image of an explosion outside the Pentagon in Arlington, Virginia, briefly went viral just after 10 a.m. local time on the morning of Monday, May 22.
The original image was shared by a “verified” account posing as a Bloomberg News feed and was then reposted by dozens of other blue-tick Twitter accounts – one of which was the ‘RT’ Russian propaganda channel.
Prime example of the dangers in the pay-to-verify system: This account, which tweeted a (very likely AI-generated) photo of a (fake) story about an explosion at the Pentagon, looks at first glance like a legit Bloomberg news feed. pic.twitter.com/SThErCln0p— Andy Campbell (@AndyBCampbell) May 22, 2023
Within an hour of it appearing, government officials and the Arlington police stepped in to clarify that the image was a fake: “There is NO explosion or incident taking place at or near the Pentagon reservation, and there is no immediate danger or hazards to the public.”
The rebuttal was not made in time to prevent a brief drop in the stock-market.
Even though some accounts quickly apologized or deleted the posts, it continued to be shared by accounts for several hours.
This is not the first time that misinformation has spread from verified accounts, some of which have up to a million followers, since Elon Musk took ownership of the platform in October 2022. These verification marks were once viewed as a guarantee of authenticity, but now seems just confirmation the tweeter has paid $8 a month to get a premium account.
On May 5, shortly after the alleged drone attack on the Kremlin, several Twitter accounts shared the false information that Russian military jets had been armed with nuclear payloads aimed at Kyiv, the capital of Ukraine.
Pope in a puffer jacket
Images generated by the AI tool Midjourney
People have already been fooled, on Twitter and other social platforms, by AI-generated images which appear to be real photographs. This includes images that went viral of the Pope wearing a puffer jacket, in early April; former US President Donald Trump being arrested, and even images of Russian President Putin greeting Chinese President Xi Jinping on one knee and kissing his hand, both on March 20. The scary thing is that it seems many users believed all three were real events.
AI Image of ‘the arrest’ of Donald Trump
AI created by Eliot Higgins
Artificial intelligence engines have been growing exponentially in recent years, prompting governments to consider legislation to control their use on social media such as forcing creators to label AI images as such.
The Biden Administration earlier this month unveiled plans to create national research institutes to evaluate AI technology and compel developers to behave responsibly in the light of potential “cybersecurity, biosecurity and safety” risks.
You can also highlight the text and press Ctrl + Enter