How AI affects your Facebook feed: Meta provides insights

The big social media player Meta wants to provide new insights into how its recommendation engine works across different products. It also plans to introduce a new option called “Why am I seeing this?” for Instagram and Facebook Reels, as well as for the Instagram Explore page.

That’s what Nick Clegg, the company’s president of global affairs, published in one blog post, The company therefore wants to make its AI systems more transparent and give users more control over the content displayed. “These systems make it more likely that the posts you see are relevant and interesting to you,” Clegg wrote.

These are the new features on Instagram and Facebook

Meta is currently testing a new system for Instagram to mark posts as “Interesting” to see more content of this type. Users can already tap the three-dot menu on the post and tap “Not Interested” to see fewer posts of this type.

The company also said it is also working on making Facebook’s “show more, show less” controls “more prominent”. Currently, they can be accessed via the three-dot menu on a featured reel or post. It’s not yet clear if Meta will move those controls elsewhere.

The secret behind the 22 system cards

It was also made public that users could now read through 22 different “system cards” that should bring more clarity about the types of AI systems used in different areas of the apps. That writes the technology page Gizmodo. It also aims to show types of user data that they rely on to function.

The extensive system maps aim to give the public one of the best glimpses yet of how Metas models work, and come after years of criticism from researchers and advocates who have criticized the company’s commitment to transparency.

Gizmodo already had insights into the system maps on Thursday. They aim to provide detailed insights into how AI systems are delivering content across a variety of meta-services, from Facebook’s Feed and Marketplace to Instagram’s Stories and Explore, and more than a dozen in between.

The failings of Meta

Each of these cards consists of four sections that provide an overview of the AI ​​system used, a summary of how the AI ​​system works, a section about the customizable controls available to users, and another section with more details about the AI ​​system used, which ultimately delivers the content.

Clegg mentioned in the blog post that the company shares some information about signals — like liking or sharing a post — that influence recommendations in the Transparency Center. However, he pointed out that the company had failed to “disclose signals that could make it easier for people to circumvent our protections.”

In parallel, Meta has discussed in his AI blog the possibility of developing AI models with parameters on the order of “tens of trillions” – much larger than current models like ChatGPT and GPT-4.

Almost finished!

Please click on the link in the confirmation email to complete your registration.

Would you like more information about the newsletter? Find out more now

Leave a Reply

Your email address will not be published. Required fields are marked *