Meta AI Support Assistant Rolls Out on Facebook, Instagram for Help With Account, Moderation Issues

Meta Platforms announced the rollout of the Meta AI Support Assistant on Facebook and Instagram on Thursday. This move is part of the Menlo Park-based tech giant's ambitions to introduce tools for support and content enforcement on its apps. The new artificial intelligence (AI) assistant is claimed to be able to respond to questions and take action on behalf of the user. Meta also claimed positive results of the changes it introduced last year to focus its proactive enforcement towards illicit content on its platform.

AI Support Assistant on Facebook, Instagram

In a blog post, Meta said that it is currently rolling out its AI support assistant globally on Facebook and Instagram across Android, iOS, and desktop-based Help Centres. It provides round-the-clock assistance for a wide range of account-related issues, with a claimed response time of under five seconds. As explained by Meta, the Meta AI Support Assistant can provide information, answer questions, and take actions on a user's behalf. These include reporting scam or impersonation accounts, helping users understand why their content was removed and providing assistance in appealing removals, managing user privacy, changing passwords, and editing user profiles.

These actions are currently limited to Facebook, but will be enabled on Instagram, too, in the future. The company said that the assistant is integrated directly into its apps, making it easier for Meta to provide support. It is available in all languages supported by Facebook and Instagram.

Apart from this, Meta is experimenting with more sophisticated AI systems to improve content moderation on its platforms. These are claimed to help identify content violations such as scams, frauds, impersonations, and harmful content more effectively and minimise false positives in the process.

The company claimed early testing has shown promising results, with the new AI systems being able to detect and mitigate around 5,000 scam attempts per day. In comparison, these were said to have gone previously unnoticed. Additionally, they are claimed to have helped reduce reports of celebrity impersonation by over 80 percent, while also identifying twice as much violating adult solicitation content compared to human review teams.

As per Meta, its AI systems can operate across languages spoken by 98 percent of internet users, a significant improvement over its earlier coverage.

While AI is being given more responsibility with regard to content enforcement, the company emphasised that human oversight remains critical. It plans to gradually deploy these AI systems over the coming years, but the responsibility of handling complex and high-risk decisions will still be with human reviewers, as AI focuses on large-scale, repetitive tasks.