In a blog post, Nick Clegg, Meta’s President of Global Affairs, emphasized the company’s commitment to transparency and user empowerment.
Additionally, Meta is testing a new feature called “Why am I seeing this?” that allows users to understand the reasons behind content recommendations.
The company is also introducing the “Interested” and “Not Interested” options to tailor content suggestions further. Meta has released 22 “System Cards” that explain various AI systems used across Facebook and Instagram.
These cards delve into topics such as Facebook Feed recommendations, Facebook Group timelines, suggested people and groups, Instagram Reels recommendations, notifications, and Stories AI.
While the information provided requires some time to comprehend fully, it represents an important step towards transparency.
Moreover, Meta acknowledges the impact of signals such as liking and sharing posts on user recommendations and has made some of this information available in its Transparency Center.
However, the company has refrained from disclosing signals that might be exploited to circumvent platform defenses.
Clegg recognizes that users’ actions, in combination with algorithms, influence content display. Yet, users often lack knowledge of the signals captured to train AI-powered feeds fully.
In related news, Meta has discussed the potential development of AI models with tens of trillions of parameters, surpassing the size of current models like ChatGPT and GPT-4. Such large-scale models can provide profound insights into user behavior.
However, this increased capacity may render existing user controls insufficient to manage content recommendations effectively.
Clegg previously acknowledged that user actions play a vital role, but individuals remain unaware of the complete range of signals captured by the platform.
To foster greater understanding and external scrutiny, Meta has introduced a Content Library and API for researchers.
This resource provides access to public posts, pages, groups, and events from Facebook, as well as public posts and data from creators and business accounts on Instagram.
In recent years, Meta has faced scrutiny regarding its content recommendation systems. Allegations of negative effects on political and news content, as well as concerns about Instagram’s impact on teenagers, have highlighted the need for change.
Meta’s CEO, Mark Zuckerberg, revealed that AI-powered Reels suggestions have led to a 24% increase in time spent on Instagram. As part of its efforts to appease regulators and privacy advocates, Meta aims to fill 30% of users’ feeds with algorithmic recommendations.