Facebook's parent company joins a growing list of big tech companies delaying or avoiding product launches in the EU, citing tough and unfavorable EU regulations.
Mehta said it will not release its planned multimodal Llama AI model in the EU due to the EU's “unpredictable” regulatory environment.
A multimodal AI model is one that can process multiple formats, including text, images, audio, video, etc. Meta launched its latest large-scale language model (LLM), Llama 3, in April, and announced at the time that a multimodal version would be available in the coming months.
The company, which is currently racing to build cutting-edge generative AI models, says its latest technology will never see the light of day in the EU.
“We plan to release the multimodal Llama model in the coming months, but will not release it in the EU due to the unpredictable regulatory environment there,” Mehta said in a statement.
Axios reported that Meta was concerned about how to train its models using EU data while remaining GDPR compliant: Last month, the company backed away from plans to train large language models (LLMs) using public content shared by adults on Facebook and Instagram after extensive discussions with the Irish Data Protection Commission.
Privacy advocacy group Noyb has raised serious concerns about Meta's plans, arguing that its intention to use AI training materials derived from publicly available and licensed data that may contain personal information violates GDPR.
Llama 3
When Meta announced Llama 3, it claimed the model showed “state-of-the-art” performance on various industry benchmarks and included new features such as “improved inference.”
Meta teased the power of Llama 3 earlier this year, when the company announced that it was using two “data center scale” clusters with over 24,000 Nvidia H100 GPUs each for Llama 3's development.
Meta claimed that these models are a “big leap” from Llama 2, and that the company's goal is to create open models that are “on par with the best proprietary models available today.”
Apple Intelligence
Meta is not the only big tech company to decide to postpone a major product launch in the EU market due to strict regulations: Apple recently unveiled its own AI technology, called Apple Intelligence, which it plans to launch in the US later this year.
The iPhone maker cited similar reasons as Meta, saying Apple Intelligence isn't scheduled to launch in the EU until 2025.
While Meta's EU challenge is based on GDPR, Apple has said it has issues with the requirements of the Digital Markets Act (DMA), which aims to promote competition and prevent anti-competitive behaviour within the EU.
The DMA requires companies to make their products interoperable with those of their competitors, which Apple argues puts user privacy and security at risk.
EU leaders aren't convinced. “It's very interesting that they say, 'We're going to introduce AI where there is no obligation to enable competition,'” EU competition official Magrete Vestager said recently when asked for her thoughts on Apple's delay.
“I think this is the most astonishing, open declaration that they are 100% aware that this is just another way to cripple competition where they already have a presence.”
Meta launched its X-rival app, Threads, in July last year and made it available in most major markets, including the UK, US and India. However, the company held back on releasing the app, citing strict EU regulations. It was finally released in the EU in December.
Find out how emerging technology trends will reshape the future in our new podcast, Future Human: The Series. Listen now. Spotifyupon apple Or wherever you listen to podcasts.
