Donald Trump's recent accusation that Vice President Kamala Harris used artificial intelligence (AI) to fabricate images of a crowd of supporters is a stark reminder of AI's potential to undermine trust in reality. On Sunday, Trump posted on Truth Social, claiming that photos of Harris greeting a large crowd on a Detroit runway were doctored using AI. He asserted, "There was nobody at the plane, and she 'A.I.'d' it and showed a massive 'crowd' of so-called followers, BUT THEY DIDN'T EXIST!" However, many witnesses confirmed the crowd's presence and shared their photos from the event, contradicting Trump’s claims.
Trump’s obsession with crowd sizes, both his own and those of his rivals, is well-known. This time, he pointed out the absence of reflections on the metallic sides of Harris's plane as evidence, overlooking that the aircraft's curved surfaces and angle would naturally distort reflections. While AI has introduced new image manipulation capabilities, traditional tools like Photoshop have long enabled photo editing. AI-generated images today can often be identified as fake, but as technology advances, distinguishing real from counterfeit will become increasingly complex, especially with audio and video.
Warnings about deep fakes have heightened public awareness, but they have also made it easier to doubt any evidence one wishes to reject. The use of AI as a defence, known as the "deep fake defence," has already surfaced in legal cases, including those involving some January 6 defendants who claimed that photos of them at the Capitol were AI-generated fakes.
The broader concern is that AI's existence and the spread of AI-generated misinformation contribute to a phenomenon known as the "liar's dividend." This concept, highlighted in a Freedom House report, describes how the pervasive fear of falsehoods can lead to rejecting factual statements. As Yale historian Timothy Snyder notes, authoritarian leaders thrive in an environment where no one trusts anything, allowing them to distort reality and erode democratic processes.
Comentarios