Meta reveals to users how AI works from its algorithms on Facebook and Instagram
Meta (parent company of Facebook, Instagram, Messenger and WhatsApp) on Thursday unveiled a comprehensive set of policies and explanations about the use of artificial intelligence (AI) in its social networks and algorithms.
The tech giant announced today that they are launching 22 “system cards” for Facebook and Instagram and that these tools give insights into how AI systems classify the content a user sees, makes predictions to determine what content might be most relevant to the consumer, and also provide the user with an opportunity to personalize their experience.
“We will be expanding our ‘Why am I watching this’ button feature on Instagram and Facebook Reels (videos) in the coming weeks, (…) The user will be able to click to see more information about how their previous activity may have informed the machine learning models that shape and deliver the content they see,” explained Meta’s president of global affairs, Nick Clegg, in a statement.
“With the rapid advances being made with powerful technologies like generative AI, people are understandably excited about the possibilities as much as they are concerned about the risks. We believe the best way to address those concerns is with transparency,” noted the Meta representative.
Clegg detailed that this action is “part of a broader spirit of openness, transparency and accountability” at Meta.
To increase its transparency, Meta also announced that it will begin implementing a new set of tools for researchers in the coming weeks.
Among the data shared is a library of content that researchers will be able to request access to.
Meta says these tools will provide “the most comprehensive access to publicly available content on Facebook and Instagram of any research tool to date,” as well as help the company meet its transparency and data sharing compliance obligations.
On several occasions, Meta has been criticized for not giving researchers sufficient access to its data.
“Our ambition is to continue to be transparent as we make more AI models openly available in the future,” Clegg added.
In February, Meta released its large language model LLaMA under an open source license that allows anyone to view and edit the code.