Earlier today, claims began to spread widely on Twitter that an explosion took place near the Pentagon. Accompanied by these claims was a single photo purportedly showing smoking rising near a building, however, this photo was fake and a key example of weaponized AI generated imagery.
We can conclude from a thorough image analysis that the image is false and likely AI developed. A closer look at the image reveals that a number of the objects are out of place and clip through the background.
The lamp post that clips through the lower part of the fencing and into it, which is the best illustration of this. As you progress from left to right, the fencing itself becomes a distorted tangle of lines and loses quality. Similar to this, if you look at the building, you can see the places where the image was not fully formed, leading to significant inconsistencies.
This was also backed by photo forensics of the picture, which shows discrepancies in the pixels.
It’s from Facebook over an hour ago….https://t.co/CYGeBPNENJ
— Royal Intel ? (@RoyalIntel_) May 22, 2023
As with most developing stories, one tries to look for other photos or videos to confirm something is happening. Not only was the post unable to be geolocated (as it was AI generated), there was no other evidence of the alleged blast through open source. To top it all off, we reached out to people we know from the area who confirmed there was no smoke or presence of emergency services.
@PFPAOfficial and the ACFD are aware of a social media report circulating online about an explosion near the Pentagon. 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. pic.twitter.com/uznY0s7deL
— Arlington Fire & EMS (@ArlingtonVaFD) May 22, 2023
The post appeared to have originated from a Qanon page on Facebook, but when it hit Twitter, it was quickly spread by bots and “OSINT” accounts known for spreading unverified information. Due to the success of many OSINT accounts on Twitter in covering the war in Ukraine, many mainstream outlets have begun to rely on open source reporting to get a jump on stories. With a large influx of OSINT accounts and fake “verified” accounts of more reputable news outlets spreading “initial reports” of the alleged blast, mainstream outlets began to run headlines without verifying the information for themselves. Twitter accounts for the Russian state-media outlet RT, as well as several other foreign media services, began to tweet out the claims, spreading it even further.
This morning, an AI generated image of an explosion at the US Pentagon surfaced.
With multiple news sources reporting it as real, the S&P 500 fell 30 points in minutes.
This resulted in a $500 billion market cap swing on a fake image.
It then rebounded once the image was… pic.twitter.com/DpHgflkMXP
— The Kobeissi Letter (@KobeissiLetter) May 22, 2023
The claim was eventually confirmed to be false by local authorities and officials, however, there was a solid 45 minutes to an hour that this claim spread uncontrolled on social media. By now, the damage was done and fear was spread, not only amongst the public, but also the markets. The S&P 500 briefly fell at least 30 points after the post hit Twitter, resulting in a $500 billion market cap swing.
A fake Bloomberg account…
Tweeted an AI Generated image of the Pentagon and stated reports of a "Major Explosion."
30 handle downside move followed…
Then the tweet was deleted.
Welcome. To the Future.
This is DANGEROUS. pic.twitter.com/B9Op30JmvE
— Mr. Wayde (@THE__NINES) May 22, 2023
Coming from someone who covers news and conflicts on a daily basis, this was the first time I have ever seen an AI-generated disinformation go viral to the extent that this one did. As AI-generated content gets more mainstream, we will see more instances of this. While AI content can be fun, such as memes of Biden and Trump playing Minecraft, they can easily be weaponized. This is why it is important to not only triple check everything you read online, but also teach yourself ways to detect AI generated content through open source tools and try to geolocate content if possible. We all live in an information environment that has shifted into a digital war zone meant to mislead people and shift perspectives. This article is a cautionary tale for what is to come.