It seems that even though the internet is increasing Drowning in fake imagesAt least we can assess humanity's ability to sniff out BS when it matters. A number of recent studies suggest that AI-based misinformation has yet to have a significant impact on this year's global elections. Because it's still not very good
Over the years There is serious concern that more and more immersive content is synthetic. It may influence the audience in a dangerous way. The rise of generative AI raises fears again. This is because technology has made it much easier for anyone to create fake audiovisual material that looks real. Back in August Political consultants use AI to President Biden's voice spoof for a robocall telling voters in New Hampshire to stay home during the Democratic primary
Tools like ElevenLabs make it possible to deliver short audio commentary. of the person speaking It then simulates their voice to say what the user wants. Many commercial AI tools, though, have guardrails in place to prevent this use. But there are open source models available.
Despite these advances Financial Times In a new story, we look back on that year and find that there was little synthetic political content circulating around the world.
It refers to a report from the Alan Turing Institute, which found that just 27 pieces of AI-generated content went viral during the summer's European elections. The report concludes that there is no evidence that elections have been affected by AI disinformation because “the majority of exposure is concentrated among a minority of users whose political beliefs align with the ideological narratives embedded in the content. As mentioned,” in other words Among the few people who saw the content (before it was flagged) and ready to believe that content The content reinforces those beliefs about the candidate. Even though those who disclosed the content knew it was AI-generated, the report cites an example of an AI-generated image showing Kamala Harris addressing a rally standing in front of a Soviet flag.
In the United States, the News Literacy Project identified more than 1,000 examples of presidential election misinformation, but only 6% used AI in only when a new imaging model is released. Not during election time.
What's interesting is that it seems that users on social media are more likely to misidentify. TRUE More AI-generated images than any other method But in general Users express a high level of skepticism. And fake media can still be debunked through official communication channels. Or through other methods such as Google reverse image search.
It is difficult to quantify with any certainty how many people have been influenced by deepfakes, but the finding that they are ineffective makes sense. Today, AI images are everywhere, but images created using artificial intelligence are still It has an unfavorable quality. It shows signs of being counterfeit. The arms may be unusually long. or the face is not properly reflected on the glass surface. There are many subtle signs that an image is synthetic. Photoshop can be used to create very convincing spoofs. But doing so takes skill.
AI proponents aren't necessarily cheering this news. It means that the image created still has a way to go. Who has gone to check? OpenAI's Sora model Know that the videos produced are not very good. It looks almost identical to something created by a video game graphics engine (Speculation is that it was trained on video games.) One thing is that properties such as physics are not clearly understood.
All of this said, there are still concerns that must be raised. Report of the Alan Turing Institute do Ultimately, it is concluded that beliefs can be reinforced with realistic deepfakes that contain misinformation. Even if the audience knows that the media doesn't exist. Confusion about whether a piece of media actually causes damage or not Trust in online resources and AI imagery is already being used. Targeting female politicians with pornographic imagesThis can be damaging both emotionally and to your professional reputation. Because it reinforces beliefs about sexism.
Technology will continue to improve. So it's something to keep an eye on.