Wrote: Festim Rizanaj
The Facebook page "Lajme Nga Bota" has published today, March 10, an image claiming to show Iran's attacks on Tel Aviv, in retaliation for an American-Israeli attack on Tehran the day before.
After verifying the image on several specialized platforms for detecting content created with artificial intelligence (AI), it turned out that the photo was generated by AI.
Image verification
Several specialized platforms for verifying digital content confirmed that the image was created with artificial intelligence.
Hive Moderation suggested that the image was generated with artificial intelligence with 99.6% certainty, while Imagewhisperer.com also reported that the image was generated with artificial intelligence, with very high confidence.
On the Sightengine platform, the photo was rated as "most likely created by artificial intelligence" with 99% certainty, while analysis on InVID showed an 81% chance of being generated by artificial intelligence.
Visual analysis revealed several signs that support this conclusion: the light trails, which are supposed to be rockets or projectiles, are extremely vertical and thin, not matching the physics of light in real long-exposure photographs; the smoke appears repetitive and has a “soft” texture, unlike the chaotic smoke of real life; the building in the foreground has minor architectural deformations.
Also, the view that appears on the phone's screen does not match the scene that is supposed to be filmed, a common error in AI-generated images.
Analyze:
The image published by the Facebook account "Lajme Nga Bota" on March 10, which claims to show Iran's attacks on Tel Aviv in retaliation for a US-Israeli attack on Tehran, is generated with artificial intelligence.
This is suggested by verifications on specialized platforms, as well as some visual distortions and unusual elements in the image.
Based on its work methodology, hibrid.info considers this claim to be "fraud".
Reasoning:
The “Deception” rating is given to informational content that is manipulated in terms of visual (deepfake, cheapfake), audio, or textual content to deceive the audience. This form of deception is usually used to produce false and manipulated content.