The audios of Santos Cerdán
It is very strange that Santos Cerdán, to defend himself and avoid his resignation, has not used the wildcard of the so-called deepfake, which is already common in courts and will unfortunately become more common. The reason may be that the interested party is unaware of this line of defense, or that the audio is not a deepfake and that the corruption under investigation is real, and that he opted for the most intelligent option in these cases: to resign from politics and defend himself like any other citizen. It's also possible that he didn't want to further damage his party politically, whether the conversation is true or false, which Santos Cerdán himself probably doesn't remember, because the alleged conversation dates back six years. Be that as it may, courts are often mistaken when they believe, as almost everyone does, that human memory is like an audio and video recording. In reality, it is very precarious and, under minimally acceptable conditions, lasts no more than 48 hours. After that time, witness psychology has scientifically confirmed that memory is reconstructed and filled with false memories to fill in its gaps, producing unreal memories that the subject firmly believes correspond to the truth. A disaster that makes interrogations in court proceedings, despite their sometimes spectacular theatricality, actually serve little more than to waste time or to appeal to the judge's social prejudices so that he decides the case not in favor of the truth, but in favor of the interests of a lawyer or prosecutor.
But returning to the subject of the deepfakesIt's important to explain that these are documents, with any content, that simulate reality, often so perfectly that the voices and images heard and seen seem real. People have generally seen them in jokes or memes showing famous people saying or doing absurd things they would never say or do, at least not in public. And, what's more, they're not that difficult to make anymore. The most common generative artificial intelligence tools can produce them, some particularly well, and there are many easily found programs on the web that, in the hands of more experienced people, do them even better.
At the moment they are a entertainment In most cases, they are now beginning to be seen in the reality of the courts. They usually entered the courtroom through complaints—with an uncertain future—when they simulated that minors were engaging in explicit sexual behavior that they never engaged in, with the resulting damage to their psychological integrity. But we will increasingly begin to see written documents that appear real but are completely false, or audio recordings of conversations that never happened. Or images, in video or photographic format, that also appear real but are not. It is reasonable to assume that these falsifications can be easily detected, and often they are because they are of low quality or what they show is surreal. But there is one important fact that is very worrying: the vast majority of expert evidence used to prove the existence of a deepfake They have a low scientific quality that also diminishes at the same rate as the production techniques of the deepfakes. Which means that, currently, around 15-20% of the deepfakes are not detectable. This is a real tragedy for the future of evidence in judicial proceedings, which will sooner or later require a global rethinking of the courts' activities. There will always be deepfakes undetectable.
This is what made the Elysée Palace initially declare a few days ago that the hand that hit Macron's face inside an airplane was not that of his wife, but that it had never happened because it was a deepfake –a version they later had to correct–. The judge in Jenni Hermoso's case must have had the same thought when he explicitly rejected a TikTok video of the image we all saw because it wasn't validated, simply because. As I said, many more cases will come.
I don't know if Santos Cerdán will also use that argument in his defense. After all, the audio comes from police officers who undoubtedly have experience—limited, however—in matters of deepfakesIf it were proven that everything was the result of manipulation, and not an actual recording by Mr. Koldo, the impact would be brutal. deepfake in many future cases. The fictional world that may emerge is unimaginable for now.