Meta Platforms Inc., the parent company of Facebook and Instagram, has announced plans to utilize data from public posts, comments, and interactions with its AI from adult users in the European Union (EU) to train its artificial intelligence models. This decision comes just a month after the company officially launched AI technology in Europe, marking a significant step in its ongoing efforts to enhance its AI capabilities.
The deployment plan, initially unveiled in June 2024, faced delays of nearly a year due to concerns regarding privacy and data protection. While Meta AI has been operational in the United States since 2023, its rollout in Europe has encountered numerous obstacles, primarily stemming from stringent EU regulations on privacy and transparency.
According to Meta, users of Facebook and Instagram in Europe will start receiving clear notifications regarding the types of data that will be utilized in the AI training process. These notifications will be delivered through in-app alerts and emails, allowing users to understand how their data will be employed. Additionally, users will have the option to opt out of data sharing through an online form provided by Meta.
Specifically, the data collection will encompass user queries, questions directed to Meta AI, and public posts and comments from users aged 18 and older on Facebook and Instagram. Importantly, Meta has assured that private messages and public data from accounts belonging to users under 18 will not be included in the AI training process.
The Data Protection Commission (DPC) of Ireland previously requested a halt to Meta's data collection plans from social media posts, leading the company to temporarily suspend its AI deployment in the EU in June 2024. This decision was influenced by backlash from NOYB, a privacy advocacy organization, which called for European authorities to intervene and prevent the use of personal content on social media for AI purposes.
Meanwhile, other tech giants are also under scrutiny. Elon Musk's social network X is facing an investigation regarding the use of EU user personal data to train its Grok AI system. Similarly, Alphabet's Google has been investigated by the DPC since September to determine if it is adequately protecting user data before utilizing it in AI model development.
In a related development, Apple Inc. has announced plans to begin analyzing data on customer devices to enhance its artificial intelligence platform. The company aims to ensure user information is protected while remaining competitive in the rapidly evolving AI landscape.
Currently, Apple primarily trains its AI models using synthetic data, which simulates real-world input without containing personal details. However, this method does not always accurately reflect actual customer data. The new technology will involve comparing synthetic data with recent user email samples from the iPhone, iPad, and Mac email applications. This process will help identify which aspects of the synthetic dataset best align with real messages.
Insights gathered from this comparison will improve text-related features within the Apple Intelligence platform. These enhancements include notification summaries, the ability to synthesize ideas in Writing Tools, and user message summaries. The new system is set to be deployed in the upcoming beta version of iOS and iPadOS 18.5 and macOS 15.5, with the second beta already provided to developers on April 14, 2025.
Apple is also implementing privacy protection methods to enhance models for various Apple Intelligence features, such as Image Playground, Image Wand, Memories Creation, and Visual Intelligence. The company has employed "differential privacy" technology to refine its Genmoji feature, which helps identify popular prompts while ensuring that unique or rare prompts remain undiscovered.
These features will only be accessible to users who have opted to participate in device analysis and product improvement initiatives, managed through the Privacy and Security tab in the Settings app. This approach reflects Apple's commitment to balancing user privacy with the need for enhanced AI capabilities.
In March, Apple underwent a leadership reorganization within its AI team, which included removing Siri responsibilities from director John Giannandrea and assigning the voice assistant team to Mike Rockwell and Craig Federighi. This restructuring indicates Apple's ongoing efforts to refine its AI strategy amidst challenges in the sector.
As both Meta and Apple navigate the complex landscape of AI development and user privacy, the implications of their strategies will be closely monitored by regulators and consumers alike. The evolving regulatory environment in the EU, combined with the competitive pressures in the AI industry, will likely shape the future of how these companies approach data usage and privacy protection.
