Meta has decided not to launch its upcoming multimodal artificial intelligence models in the European Union (EU), citing the "unpredictable nature of the European regulatory environment." This decision will affect both users and businesses in the region. The tech company had planned to present its multimodal Llama model, which has the ability to process and integrate data across various formats such as video, audio, images, and text. However, due to stringent regulations regarding the use of AI and user data in the EU, Meta has opted not to launch these models in this region for now.
Impact on European Businesses and Users
Meta's decision could have several repercussions. European businesses will not be able to access this advanced multimodal technology for their work processes, despite Meta usually offering these models under an open license. Additionally, European users and companies will not be able to develop products based on Meta's new models. This may also affect companies outside the EU that use these models, limiting their ability to offer services in Europe.
Despite the imminent entry into force of the Artificial Intelligence Regulation on August 1 of this year, Meta already faces challenges with European regulations due to its need to use user data to train its AI models. The EU opposes the use of personal data by companies like Meta, backed by the General Data Protection Regulation (GDPR), which aims to give users more control over their personal information and the type of advertising they receive. In May of this year, Meta announced its intention to use public information from Facebook and Instagram to train its AI models. However, it later offered an option for users to prevent the use of their data through an objection form.
Regulatory Challenges and Repercussions
In June, Ireland's Data Protection Commission (DPC) reported that Meta had halted plans to train AI models with publicly shared content by adults in the EU. Meta expressed its "disappointment" over having to delay the training of its large language models (LLM) using public content from its platforms. In its blog, the company stated that this represents a setback for European innovation and competition in AI development, as well as a delay in bringing the benefits of AI to people in Europe.
Other tech companies have also made similar decisions. Apple, for instance, has delayed the implementation of its AI features, Apple Intelligence, and iPhone Mirroring for Mac and SharePlay Screen Sharing in the EU, due to "regulatory uncertainties" related to the Digital Markets Act (DMA).
Despite these challenges, Meta has announced plans to launch a larger version of its Llama 3 model, which operates solely with text and will be available to customers and businesses in the EU. Additionally, the company confirmed to Axios that its upcoming multimodal AI models will be launched in the United Kingdom, where, despite having a law similar to the GDPR, they do not face the same level of regulatory uncertainty as in the EU.