Meta AI development halted: data usage raises concerns

In the European Union, the company announced that will not release its next multimodal AI models due to regulatory uncertainty. One of the affected models is Multimodal Llama, which is designed to handle data in multiple formats, such as video, audio, images, and text.

Meta explained that the decision was in line with new regulations that affect the use of artificial intelligence and user data in the region, according to Axios. The company described the situation in the European regulatory environment as “unpredictable.”

In Brazil, the National Authority for Data Protection (ANPD) has also imposed restrictions on Meta’s artificial intelligence training project from early July 2024. The measure includes invalidating Meta’s new privacy policy, which allowed the use of personal data to train generative AI systems.

ANPD justifies its decision by concerns about possible harm to the fundamental rights of data ownersimposing fines in case of non-compliance. Meta will need to adjust its privacy policy to exclude the processing of personal data for the purposes of AI training, and issue an official statement confirming the suspension of this processing.

Meta expressed disappointment over the decisions, arguing that they represent an obstacle to innovation and delay the benefits of AI for EU and Brazilian citizens.

The Company ensures that its approach complies with Brazilian privacy laws and regulations.highlighting its transparency compared to other industry players who have used publicly available content to train AI models.

The activist group European Centre for Digital Rights (NOYB) filed several complaints with Meta when it announced it was going to train your AI model on data from Facebook and Instagram users in the European Union.

An Austrian NGO has expressed concern about Meta’s intention to use the data to develop artificial intelligence technology, as its characteristics have not yet been fully defined.

In addition, they stressed that The company did not specify which third parties it plans to share such user information with.

When Meta announced that this particular project was going to be stopped in Europe, NOYB said: “Meta has every opportunity to implement AI based on genuine consent, but it just chooses not to.”

In Brazil, the country’s Data Protection Authority also noted that users face excessive and unjustified barriers in obtaining information and exercising their right to object to the processing of their personal data.

The Brazilian organization has determined that the Meta project is a “risk of serious damage that would be difficult for the user to repair“.

In early July 2024, the organization specifically published the following: “Preventive measure determining the immediate suspension in Brazil of the new privacy policy of Meta, which allows the use of personal data published on its platforms for the training of artificial intelligence systems.”

Besides, Meta is subject to a daily fine of R$50,000.00 (fifty thousand reais) in the event of failure to comply with the imposed preventive measure.

In a statement to The Associated Press after the pre-emptive suspension was announced, Meta said its policies comply with privacy laws and regulations in Brazil.

They also said the decision represents a setback for innovation, competition in AI development and will lead to further delays in bringing the benefits of AI to the Brazilian population.