Meta Platforms Inc. has ambitious plans to start the training of its new artificial intelligence models. This will all be done using only public content and user interactions within the European Union, implementing as soon as this week! This latest move reflects the company’s long-standing pledge to pivot as rules and regulations evolve. It strengthens its artificial intelligence capabilities. This decision breaks new ground in the pattern set by the Goliaths of tech, such as Google and OpenAI. These include their use of data from European users to train their AI systems.
Of course, Meta wants us to know that they are deeply committed to protecting user privacy. In response, the company stated that it does not use private messages or data from users under 18 in the EU to train its models. Yet, the company seems to be making a good faith effort to comply with various data protection regulations. These rules are currently enforced by the Data Protection Commission (DPC) in the EU, acting on behalf of dozens of data protection authorities across the bloc.
Aisha, a consumer news reporter at TechCrunch, highlighted that Meta’s approach includes honoring all existing objection forms submitted by users, as well as any new ones that may arise. This would show Meta is ready to listen to their users’ concerns about how consumer data is protected and used.
“We believe we have a responsibility to build AI that’s not just available to Europeans, but is actually built for them,” – Meta spokesperson
In September 2024, Meta plans to restart efforts to use public posts from its U.K. user base for AI training. This is a big win for the public data strategy of the company and its desire to use public data, but not without complying with local laws. The company will provide notifications via in-app messages and emails to inform users in the EU about the new data usage policies.
Similarly, last year, in the face of criticism, Meta paused its plans to train large language models on public content. They achieved this all in advance of having clear legal guidance from regulators. This halt provided the company an opportunity to reflect on its practices. It further assisted in making those practices more consistent with EDPB guidelines and decisions. Yet in December, the EDPB agreed that Meta’s initial default settings were fully compliant with the law. With this ruling, the company received their green light to forge ahead.
Meta’s decision to utilize public data and user interactions for AI training reflects a broader trend among technology companies navigating complex regulatory environments. Google and OpenAI have played this strategy in the past. They feed back into their AI models directly, based on user data in the EU.
Going into the future, Meta wants to hold itself accountable and be transparent. Users will be given instructive alerts. These messages will clarify how their public engagements with Meta AI contribute to better models from the company. This program aims to build confidence in users, all while building Meta’s AI capabilities.
Leave a Reply