Md. Himel Hasnat Rafi

Fellow, Dismislab
AI image of Italian newspaper used to claim Badhon interrogated over alleged involvement with extremist org

AI image of Italian newspaper used to claim Badhon interrogated over alleged involvement with extremist org

Md. Himel Hasnat Rafi

Fellow, Dismislab

Actress Azmeri Haque Badhon was recently subjected to physical harassment and attack by supporters of the Awami League in Venice, Italy. In connection with this incident, an image purporting to be a report from the Italian newspaper Corriere della Sera has been shared on Facebook, claiming that Italian police are interrogating Badhon over possible links to an extremist organization. However, Dismislab’s investigation found that the newspaper published no such report; the image is in fact AI-generated.

Last Wednesday (9 September), the image was posted from a Facebook account. It shows a photo of Badhon and a headline beneath the masthead of Corriere della Sera which translates to, “Italian police are interrogating Azmeri Haque Badhon to verify whether she has any possible links to extremist organizations.” As of the writing of this report, the post had received nearly 1,100 reactions and had been shared 117 times. The post spread with the same claim from several other personal Facebook accounts as well (1, 2, 3).

Azmeri Haque Badhon italy attack ai altered
Screenshots of a Facebook posts sharing an AI-edited image.

To verify the claim, Dismislab first conducted a reverse image search on the photo. No trace of this report was found in any reliable source. A keyword search using the report’s headline also turned up no such report. Dismislab then searched Corriere della Sera’s website for coverage of the attack on Badhon. Two reports on the incident were found there (1, 2), but neither contained any information about a police interrogation or allegations of links to an extremist organization.

For further confirmation, Dismislab went through the newspaper’s e-paper editions from the past week. No such report about Badhon was found in any issue from this period, including the September 8 edition specifically cited in the circulating post.

A close examination of the circulated image found small text in Italian on a red background just below the newspaper’s nameplate, reading, “This is a fake news item, this is not a real event” (original: Notizia fittizia – Questa non è una notizia reale), indicating that the purported report in the image is not real news.

To check whether the image was AI-generated, Dismislab then examined it using the AI content detection tool Hive Moderation, which flagged a 63.7 percent probability that the image was AI-generated. When examined using OpenAI’s own verification tool, the image was identified as having been generated using an OpenAI tool.

Azmeri Haque Badhon italy attack ai altered
Analysis of ‘AI or Not’ and ‘OpenAI’

Meanwhile, a report by the Italian newspaper La Stampa states that Digos, a special branch of the Italian police, has opened an investigation into the attack on Badhon and reviewed security camera footage from the area. This report, too, contained no information about any interrogation related to Badhon’s alleged links to an extremist organization. Multiple Bangladeshi news outlets have also published reports on the attack on Badhon, but none of them contain any information about a police interrogation either.

Azmeri Haque Badhon italy attack ai altered
The published covers of Corriere della Sera newspaper from September 7–9.

In other words, the image circulated in the name of Corriere della Sera, claiming that Italian police are interrogating Badhon over alleged links to an extremist organization, is AI-generated.