After years of working with third-party fact-checking organizations, Meta has announced the shutdown of its fact-checking program in the United States. This move marks a major shift in how the platform handles information, turning toward a more participatory approach: the arrival of “Community Notes,” inspired by the system already used on X (formerly Twitter).
A Shift Toward a Collaborative Model
In place of traditional fact-checking, Meta plans to experiment with a system of user-generated “notes” designed to counter misinformation more directly and collaboratively. This feature will allow users to add context to public posts, flagging potentially misleading or incomplete information and supporting their comments with credible sources. Notes will only appear if a diverse group of reviewers agrees on their helpfulness, aiming to reduce bias and abuse of the system.
The decision comes at a particularly sensitive time, with the 2024 U.S. presidential elections approaching, raising questions about the effectiveness of the new approach in combating viral misinformation.
AI Integration, New Features, and Privacy
At the same time, Meta is expanding its AI assistant across all its platforms. Following its debut on WhatsApp, Meta’s intelligent assistant – Meta AI – is now available on Instagram, Facebook, and Messenger, powered by the new LLaMA3.2 language model.
The assistant can be accessed through the messaging sections of the apps: on Instagram, it appears in the direct message search bar; on Facebook, it can be launched via the main app or directly from Messenger. To use it, users must have the latest version of the app installed on Android or iOS devices.
Meta AI can answer questions, offer personalized suggestions, share motivational quotes, assist with writing, translating text, and planning activities. However, it currently does not support image generation or integration with external apps.
Full Mobile Experience, Limited Browser Use
The user experience is optimized for smartphones. When accessed via web browsers, several limitations remain: on Facebook web, for example, Meta AI chat may be unavailable, and on Instagram desktop, interaction is limited to viewing only, with no ability to send messages. Additionally, some features available on WhatsApp – like forwarding messages from other conversations – are missing.
Data Protection and AI Training
Meta has clarified that the AI assistant cannot access users’ private content. However, interactions may be used to improve the system outside of the European Union. Public content shared on Meta’s platforms can contribute to AI training. Users can delete chat histories at any time, and to permanently remove collected data, an official form must be submitted.
A Double Challenge for Meta
Between rolling out its AI assistant and scaling back fact-checking, Meta seems to be betting on increased user responsibility. However, this strategy raises complex issues around privacy, effectiveness, and controlling misinformation—especially with major events like the U.S. elections on the horizon. The success of this new phase will depend on system transparency, the quality of user-generated information, and the AI’s ability to support—rather than replace—critical thinking.
