A review of Meta's (Facebook) challenges in identifying AI-generated content and unrealistic images that circulate on social media and mislead users.
AI fakes
Fake photos and posts
Recently, posts with images created using neural networks that mislead tens of thousands of users. Some of them, despite clear signs of artificial intelligence, were not labeled as false or AI-generated. For example, a post about a robot that allegedly attracted thousands of spectators to its first concert, with absurd details such as using a non-existent Apple M3 Pro processor and inviting celebrities. Although Facebook ultimately flagged the post as false, it managed to rack up hundreds of likes, comments and shares.
AI identification issues
Meta announced plans to add "Created with AI" tags " to photos uploaded to Facebook, Instagram and Threads, but the system does not work correctly. Meta's tools sometimes mistakenly flag real-life photos as AI-generated, such as photos from an Indian Premier League cricket tournament on Instagram or a photo of a basketball game taken by former White House photographer Pete Souza.
Disadvantages of recognition algorithms
Instagram algorithms can flag images as AI-generated when photographers use Tools like Adobe's Generative Fill for removing small objects. However, this process is flawed and inconsistent. Given Meta's long-standing problems with content moderation, a similar situation may arise with artificial intelligence. Users should learn to recognize fake or artificial content to avoid being fooled by misleading social media posts.
Glossary
- Meta is a company that owns social networks Facebook, Instagram and Threads.
- Neural networks are artificial neural networks used to create images and texts using artificial intelligence.
- Generative Fill (Adobe) is a tool in Photoshop for removing objects from images using artificial intelligence.
- Pete Souza is a former White House photographer.
Links
- Facebook flag for false post
- Pete Souza basketball game photo
- PetaPixel article on AI identification errors on Instagram
Answers to questions
What are neural networks and how do they work with images?
How do users react to synthetic images on social networks?
How is Meta trying to combat the spread of synthetic images?
What are the challenges with AI image recognition on social media?
What risks are associated with the distribution of synthetic images on social networks?
Hashtags
Save a link to this article
Discussion of the topic – Facebook has lost trust: generated photos and fakes have taken over the social network
Recently, cases of dissemination of fake photos created using neural networks and fake stories that users mistake for real have become more frequent on Facebook.
Latest comments
8 comments
Write a comment
Your email address will not be published. Required fields are checked *
Юлия
I see more and more AI-generated artificial images being posted on social media. This can lead to the spread of misinformation and fake news. 🤔 The meta needs to do a better job of tracking such content and labeling it clearly.
Алёна
I agree, this is a really big problem. But users also need to be more critical and not believe everything they see on the Internet. 👀 Digital literacy needs to be developed.
Макс
These fake posts amuse me! 😄 People discuss non-existent things with such seriousness. I remember how everyone believed in “the grandmother who tied the tank.”
Григорий
Phew, there are so many fakes out there, it’s a real disaster! 😒 In my time there was no such nonsense, everything was fair. These neural networks of yours are only fooling young people.
Софья
And I believe that AI technologies open up new opportunities for creativity and self-expression. 🎨 Yes, we need to fight disinformation, but we should not reject the full potential of neural networks.
Антон
The problem is that most people don't understand these technologies. So they believe everything they see. 🤷♂️ The population needs to be better educated about AI issues.
Ксения
I once came across a fake post about a “robot’s first concert.” 🤖 It was funny to read the comments of people who believed this nonsense. But in general, such fakes are dangerous and need to be stopped.
Виктор
Let's better discuss how we can develop these technologies further and use them for good! 🚀 Neural networks are the future, and we must be at the forefront of progress.