Back

Surge in AI-Generated Image Misinformation

A recent study by Google and several fact-checking organizations has revealed a significant rise in the use of AI-generated images in misinformation. This rapid increase places AI images almost on par with traditionally manipulated images created using tools like Photoshop.

The study, which analyzed nearly 136,000 fact-checks from 1995 to November 2023, shows a notable shift starting in spring 2023, coinciding with viral fake photos such as Pope Francis in a puffer coat. Researchers noted that AI-generated content, specifically image misinformation claims, has surged, indicating a swiftly evolving landscape.

AI democratization has enabled nearly anyone to spread false information online. The study highlights that while AI-generated images are on the rise, the most prevalent form of image-based misinformation remains real images taken out of context. About 80% of fact-checked misinformation claims involve media, with video becoming increasingly prominent since 2022.

Real images, paired with false claims, continue to be a major issue. These images often require no AI or photo-editing to spread misinformation. Screenshots with misleading text also play a significant role in misinformation, as they are highly shareable and often misrepresent the context of the original content.

The researchers also emphasized that while AI-generated images have become more sophisticated, traditional signs of AI manipulation, such as distorted features or garbled text, are diminishing. This makes detecting fake images increasingly challenging.

The study warns that the evolving nature of disinformation, especially with advancing AI technologies, poses a continual challenge. As AI tools become more advanced, they may start generating realistic images like screenshots and memes, further complicating the battle against misinformation.

In combating this issue, experts like Alexios Mantzarlis, director of the Security, Trust, and Safety Initiative at Cornell Tech, suggest that people need to scrutinize the source and distribution of content rather than relying solely on the visuals.

Despite its comprehensive analysis, the study acknowledges limitations, such as focusing only on English-language misinformation and relying on fact-checked claims that might not cover the entire scope of misinformation.

The findings underscore the need for continuous adaptation and the implementation of safeguards to mitigate the harms of evolving disinformation technologies. As the landscape rapidly changes, staying vigilant and questioning the origins of online content is crucial in combating misinformation.

Source: NBC News