Facebook uses AI to recommend videos: Are Reels about to get even more addictive?

Facebook uses AI to recommend videos: Are Reels about to get even more addictive?،

Artificial intelligence is making headlines as big tech players fight for supremacy in the field. Google touts its Gemini, while Samsung displays its Galaxy AI, recently launched with the latest Galaxy S24. Apple is also on board, developing its own AI features. And let's not forget Meta, which keeps pace with the competition.

According to Engadget, Meta is shaking things up by infusing AI into Facebook's video recommendations. Facebook head Tom Alison revealed that the new AI-powered video recommendation algorithm is already in action on Reels – Meta's answer to TikTok.

The plan is to roll out this new AI-based recommendation engine to all areas of Facebook that offer videos, from Facebook's main feed to Groups. This is part of a larger “technology roadmap” which is expected to run until 2026, Alison revealed during a technology conference at Morgan Stanley in San Francisco.

Making a splash in the TikTok arena has been a key goal for Meta, especially as the Chinese app, famous for its vertical video clips and robust recommendation engine that seems to know what hooks users, has soared in popularity in the United States in recent years. . When Meta tested the new AI-based recommendation engine with Reels, watch time increased by around 8-10%, as Alison shared.

Alison highlighted that the results showed them that this new model architecture learns from data much more efficiently than the previous generation and that the company is on the “right way.”

Until now, Facebook had separate video recommendation engines for Reels, Groups, and the main Facebook feed. However, with the positive results seen in Reels, the company is now looking to apply the same AI-powered engine to all of these products. Alison said:

Instead of just powering Reels, we're working on a project to power our entire video ecosystem with this single model, and can then add our feed recommendation product to be served by this model as well. If we achieve this, not only will the recommendations be more engaging and relevant, but we believe their responsiveness can also improve.

Meta's efforts to integrate AI into all of its products have gained momentum with the explosion of AI technology following the launch of OpenAI's ChatGPT in late 2022. In a video earlier this year, Mark Zuckerberg revealed that the company was spending billions to acquire hundreds of them. of thousands of expensive NVIDIA GPUs critical to training and powering AI models.