OUR Artificial intelligence was once something that the average westerner voter can be described, but rather abstractly. The devices he used may have incorporated him to some extent into their technological functions, but he himself had no “tangible” connection with him, at least not consciously.
That all changed last year, however, when millions of “regular” voters began testing artificial intelligence programs such as OpenAI’s ChatGPT and DALL-E.
In this way, artificial intelligence as a technology is “democratized”, but in this way creates an environment of new pressure at the expense of democracy itself, as Thor Benson notes in his article for the American magazine Wired.
Photographing Trump’s arrest in anticipation of Trump’s arrest. pic.twitter.com/4D2QQfUpLZ
— Eliot Higgins (@EliotHiggins) March 20, 2023
AI tools could begin to distort reality and become a threat ahead of the 2024 US presidential election, especially if audio files, images and videos of candidates that are the product of artificial intelligence are circulated.
The line between truth and falsehood may begin to blur much more than it already does in an age of rampant disinformation, Wired notes in its article.
“We have seen some pretty dramatic changes in terms of production tools, especially over the past year. I think the scale of the content we’re seeing right now is directly related to this dramatic opening up of accessibility (i.e., AI tools to a much larger audience of regular users),” comments AI expert Henry Eider in an interview with Wired.
The question is not whether political/partisan propaganda material will come from AI programs, as this is already happening.
Eliot Higgins of the Bellingcat Research Group tweeted last March photos of the police officers who arrested Donald Trump, photos he created with the Midjourney v5 program, he publicly acknowledged, saying they were clearly not real shots.
“Obvious”, however, is not always equally obvious to everyone, especially if the motives for its creation are malicious or unjust.
It is noted that a fake video was released a few months ago in which Joe Biden states that “he may have doses of dementia here and there, but he also has a razor-sharp mind …”, making him “smarter than the greatest minds” . “technology discovered a few months ago”, while scenes of chaos of a different, Republican origin, videos of what will happen in the US and outside the US if (“what if …”) Biden secured a new four-year the White House term has also been a product of artificial intelligence.
Recall that last year a fake video was published in which President of Ukraine Volodymyr Zelensky called on his soldiers to surrender, laying down their arms.
So the question is not whether artificial intelligence will become a tool in the hands of party propaganda, as it already is.
Now the question is, what happens when anyone can open their laptop and, with minimal effort, quickly create a compelling but deepfake “document” starring a politician?
A deepfake of Ukrainian President Volodymyr Zelensky urging his soldiers to lay down their arms was reportedly uploaded to a hacked Ukrainian news site today. @Shayan86 pic.twitter.com/tXLrYECGY4
— Mikael Thalen (@MikaelThalen) March 16, 2022
Programs such as DALL-E, Midjourney and Stable Diffusion allow the user to create images using keywords. Programs such as those by ElevenLabs allow the user to create voices that imitate other voices by saying what the user dictates. It’s not easy to produce truly compelling deepfake videos at the moment, but even that could change in a year, according to AI expert Henry Eider in Wired magazine.
“I don’t think there is a website today where you can say, ‘Make me a video of Joe Biden saying this and that. It doesn’t exist now, but it will someday. It is the matter of time. Experts (including Runway, Google and Meta) are already working on converting text to video,” says Hani Farid, professor at the School of Computer Science at the University of California at Berkeley.
Thor Benson of Wired, for his part, notes that “it’s easy to imagine someone like Donald Trump sharing deepfake content on social media claiming to be real.” “Just before the election, a deepfake starring Biden could be released,” Benson continues, “and many may never know that it was created by an artificial intelligence program.” After all, research has already shown that fake news travels faster than real news.
“Even if deepfakes do not become more mainstream until the 2024 elections, which are still 18 months away, the fact that such content is created could influence the elections. “Knowing that fake images, sounds, and videos can be created relatively easily can lead people to distrust even legitimate material they encounter,” Benson notes on the Wired magazine website.
However, based on technological developments, antidotes (watermarks, fingerprints, encrypted C2PA signatures, fact-checking tools) can be created that will defeat deepfakes …
According to Wired, Ars Technica, The Verge
Source: Kathimerini
Anna White is a journalist at 247 News Reel, where she writes on world news and current events. She is known for her insightful analysis and compelling storytelling. Anna’s articles have been widely read and shared, earning her a reputation as a talented and respected journalist. She delivers in-depth and accurate understanding of the world’s most pressing issues.