What data from EU users will Meta AI use for training?

Question

Accepted Answer

Meta AI to Utilise Public European User Content for Training, With Opt-Out Options Available

Meta AI will be trained using public content and conversations from its European users, including posts on platforms like Instagram and Facebook. However, individuals within the EU have the ability to opt out of having their data used for these generative AI training purposes. WhatsApp messenger services will not be impacted by these changes, and data from users under 18, as well as private messages between family and friends, will also be excluded from the training dataset.
Public Content from Facebook and Instagram to be Used

The data designated for training Meta's AI models encompasses a wide range of public content shared by users across Meta's platforms. This includes, but is not limited to, Instagram photo captions and Facebook comments. The company stated that this training is intended to enhance the AI's comprehension of European cultures, languages, and history, thereby improving its functionality for millions of people and businesses in the region.
Opt-Out Mechanism for EU Users

Meta has implemented an opt-out process for European Union users who do not wish for their public content and interactions to be used in the training of generative AI models. The company has indicated that this opt-out form is designed to be easily accessible and user-friendly, and that all prior and future objections will be respected. This move comes as Meta seeks to compete with other AI models like ChatGPT.
Exclusions from Data Training

Certain categories of user data will be explicitly excluded from Meta AI's training regimen. Notably, the WhatsApp messenger service will remain unaffected by these data collection changes for the time being. Furthermore, the company has confirmed that data pertaining to users under the age of 18 will not be utilised, nor will the content of private messages exchanged between family members and friends.
Regulatory Context and AI Development Investment

The decision to utilise European user data for AI training follows a period of delay in the region's rollout of Meta AI, attributed to overlapping EU regulations concerning personal data, AI, and digital markets. Meta asserts that its approach mirrors that of other major technology firms like Google and OpenAI. The company is significantly investing in AI development, earmarking US$60-65 billion for the current year, primarily for infrastructure such as data centres and servers essential for building advanced AI models.

Reading history

Hide →

This history is cleared when the page is closed.

What data from EU users will Meta AI use for training?

Loading

Was this article helpful?

Helpful

Not helpful

Report an issue

2

0

0

Share with friends