Meta to train its AI with public content from EU users
Meta will start using public content, such as posts and comments on its platforms, to train its artificial intelligence (AI). User interactions with Meta AI will also be used as training fodder, the social media behemoth said in a blog post yesterday (14 April).
As the parent company to Facebook, Instagram and WhatsApp, Meta says it has a “responsibility to build AI that’s not just available to Europeans, but is actually built for them”. It said that the data it collects will train generative artificial intelligence (GenAI) models to “understand the incredible and diverse nuances and complexities” of Europeans.
Pushback and Launch in Europe
Meta faced pushback from European privacy watchdogs over its chatbot Meta AI, delaying its launch in the region by more than a year. The features had been made available to US users back in 2023.
Last year, the tech giant paused plans to train its large language models using public content shared by adults on Facebook and Instagram, following intensive discussion with the Irish Data Protection Commission (DPC). However, last month, Meta was finally able to launch its AI chatbot in Europe. Although, the DPC, in a statement to SiliconRepublic.com said that they will keep Meta AI “under review as it rolls out to users”.
User Data and Privacy
Private messages are safe from being used for training, the social media giant clarified. “Additionally, public data from the accounts of people in the EU under the age of 18 is not being used for training purposes,” it said.
Moreover, the company noted that a December opinion issued by the European Data Protection Board “affirmed” that its original approach, which received pushback, had met legal obligations.
Apple's Approach
Meanwhile, Apple says it has found a way to improve its AI models without training it on user data. In a blog post yesterday, the iPhone maker floated creating a synthetic data set “representative” of aggregate trends in real user data, without actually collecting any emails or texts from its devices. Apple said that it will be comparing the synthetic data set to a sample of recent emails and texts from users who have opted in to its Device Analytics programme.
Don’t miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic’s digest of need-to-know sci-tech news.
