It’s one audio track of a few seconds. A voice similar to that of is heard Daniela Santanchénow a former minister after her perhaps not exactly voluntary resignation: “I want to be very clear, I don’t intend to go on trial. If I jump then the whole government will jump, from Meloni to La Russa. And all the other comrades. The trials will have to expire and those who have ears to hear should hear.” TikTok’s algorithm has begun to propose it again in the last few days. The videos that are used with this audio they are very artisanal. Old photos of Santanché, newspaper screenshots. We analyzed it to understand where it comes from. Except in very unlikely cases, it is a fake.
Let’s start from audio files. We extracted it from the videos and inserted it into a series of platforms dedicated to content created with artificial intelligence. The results are almost always the same. We leave you some examples based on tests we did. A premise: the systems for understanding whether an audio was created with artificial intelligence are not perfect, especially when the AI used it’s not one of the best known ones.
- Detector24 | The audio is considered deepfake: 92.7% probability
- Undetectable | The audio is considered a deepfake: 69% chance
- Truthscan | The audio is considered a deepfake: 70% chancewith segments al 99%
- Deepfakes Detection | Here the audio is considered reliable al 95%
Let’s put it this way. Taking an average of the results obtained, we can argue that the audio of Daniela Santanché that we heard is on average recognized as a fake. Alone one case in four he sees it as original. By ear, those are also the parts most similar to the products of artificial intelligence more problematic: “If I jump then the whole government will jump, from Meloni to La Russa. And all the other comrades. The trials will have to expire and those who have ears to hear should hear.” But there’s more.
“He asked me to go out with him, but he used ChatGPT”: the risks of social offloading for young people
The sources of the audio attributed to Daniela Santanché
Audio has been circulating for some time and only moves in those that steal a definition of historiography I am secondary sources. We found it taken up by small social pages, creators who comment on the news of the daypolitically themed memes and pages that are very aligned on political areas opposite to those of Daniela Santanché. There are never direct references to sources. On the contrary. Sometimes these videos are validated by screenshots of newspaper articles. We checked those too. A video puts this audio under a title of the Fatto Quotidiano: “The prosecutors: “The Visibilia trial at risk of statute of limitations” after the hearing was postponed to September. The original article was published in July 2025. There are no references to audio, comrades or other similar content.
And again. In another video the audio is edited below the screen of the newspaper la Repubblica with a title where we read: “Santanché: whoever thinks of investigating me will end up like Paolo Borsellino and Giovanni Falcone, magistrates think of the people and don’t touch the politicians”. Needless to say, Google doesn’t give back no results looking for this title, if not this video published in the Facebook group You’re from Cupertino SE.
The video at the origin of everything
One of the first clips that started it all dates back to July 2025. This time it is not just an audio but a video of Santanché, where the former minister is dressed in an orange jacket in what appears to be a press conference. We analyzed the image: it comes from a conference of the Brothers of Italy in March 2025. We listened to the statements released by the press agencies and reviewed the videos from that day. No such statements appear, on the contrary. Santanché reiterated the government stability.
The domino effect: inside the fake audio tunnel
Despite this, the audio starts playing. Maybe also because it contains the perfect ingredients. It’s strange but the rumor is plausible. It surprises the user. Contains elements that leave you immediately understand a conspiracy and a blackmail system. That’s why it exploded. We find it reproduced in dozens of videos and commented on by creators who wink at the public. From TikTok it reaches other media. And often it is played right onaudio ambiguity. A creator posted a Reel on Facebook in which he makes a series of surprised expressions for the camera as he runs the camera.audio in the background and then ask with “If it were real audio it would be very serious, but it’s not true. Right?”
Perhaps the peak of this oversized diffusion for an audio born from a small page that uses AI to do a bit of propaganda has arrived two days ago on TikTok. Here the audio attributed to Santanché becomes the soundtrack of a clip in anime style dedicated to the clash between the former minister and the Prime Minister Giorgia Meloni: Santanché seems animated by a wind that emphasizes her aura while her eyes glow red. Then the scene shifts to a navy admiral-style Giorgia Meloni. Here, it is easier to demonstrate that we are not faced with a scene that actually happened.
