A fake explosion photo allegedly produced by artificial intelligence at the Pentagon spread rapidly on social media, causing great confusion among users and a short-lived wave of sales on the US stock market. The spread photo was shared by media outlets as if it were real, and the effect of the event grew even more. This situation once again reveals the lack of detection of artificial intelligence content.Fake Pentagon explosion
The fake photo showing smoke rising outside the Pentagon was shared by the Russian state media outlet and other accounts, along with allegations that an explosion occurred in the area. Although the Fire Department tweeted to deny the situation after the incident came to the fore, the damage had already been done.U.S. stocks briefly tumbled after the photo went viral, but the recovery was quick after the photo was revealed to be a hoax. A surge of 500 billion dollars arose due to these short-term sales transactions.
verified accounts
Twitter is a particularly favorable environment for the dissemination of fake images or videos, as this platform is not only a social network, but also a source of news for people all over the world. The fake Pentagon explosion image shared by the fake @BloombergFeed account with blue ticks looks too “realistic” to be considered real to the careless eyes.Therefore, the “blue tick” that anyone can buy on Twitter and the damage it causes comes to mind. Moreover, the display of blue ticked accounts is also encouraged by the algorithm. That’s why paid “troll”, “bot” or “fake” accounts have the power to reach more audiences than ever before. Tweets that are thought to be real spread rapidly and are seen by millions of people in just a few hours.
Details have never been more important
On the other hand, there is a part of the event that falls on the user’s side. Currently, the images produced by artificial intelligence may have reached a fascinating realism, but when you take some time and examine the images, the mistakes are very obvious. For example, the structural deformation of the fences in the Pentagon incident illustrates this. Likewise, it is obvious that the pavement looks strange.
In the light of all this, today, it is necessary to pay more attention to the realism of the content shared by approved accounts or media organizations. Artificial intelligence materials created by visuals, audio files, videos or a combination of all these can now be made easily. Although there are studies on artificial intelligence tools that detect artificial intelligence contents, there is no effective use. Perhaps holding platforms accountable for detecting fake content could speed up the process.
31909
