The growing proliferation of "deepfakes" is a cause for concern in a year in which elections are being held overseas

[rUparaHmaN014]

The problem, says the Mozilla Foundation in its study, is that technology (AI in particular) is put into circulation without being equipped from the start with measures against potentially malicious uses . Moreover, instead of using uniform standards, each of the "players" that make use of AI use a different approach when regulating this technology.

From simple text specifications, AI can generate eerily realistic photos. And videos powered by the latest technology are also imbued with astonishing realism. OpenAI, the parent company of changsha mobile phone numbers database ChatGPT, recently presented Sora , a new tool for generating videos from text prompts that are almost indistinguishable from real ones. However, in Sora, the labels that indicate the true origin of the videos are so small that they are difficult for the end user to locate, criticizes Ramak Molavi Vasse'i.

In 2024, when the US presidential election is due to take place, there is great concern that the increasingly ubiquitous “deepfakes” could influence the results. Indeed, across the pond, shortly before the Democratic Party primaries in New Hampshire, robocalls were made with the voice of President Joe Biden urging voters not to go to the polls.

Fake audio content is seen as a major problem because it is more difficult to identify than deepfakes in images and videos . In fact, AI-generated audio is already being used by criminals to carry out robberies using the voices of victims' relatives as cover.