Meta utilizará datos públicos de sus usuarios en la UE para entrenar a sus modelos de IA

Meta utilizará datos públicos de sus usuarios en la UE para entrenar a sus modelos de IA

Meta plans to train its generative AI models on public user data and interactions within the EU, marking a significant shift in how social media giants leverage open information for model development. While the company allows users to opt out, this policy highlights the complex intersection between commercial AI training and European privacy regulations such as the GDPR. The initiative underscores the critical importance of data transparency and user consent in the era of generative AI. For the open data community, this scenario illustrates the tension between utilizing publicly available digital footprints for technological advancement and protecting individual privacy rights under strict legal frameworks. Ultimately, this decision challenges the notion that "public" data can be freely used for AI without explicit permission. It emphasizes the need for clearer standards on data provenance and the ethical implications of training models on personal conversations, reinforcing the demand for robust governance in open data practices.

Related news

• Apple needs its AI to be better but wants you to know it still really, really cares about your privacy
• On-device Apple Intelligence training methods seem to be based on a controversial technology
• How Apple gets AI training data without violating anyone’s privacy
• Apple Uses New Tech That Compares Synthetic Data With Real Emails To Train AI Models, Then Applies Embeddings And Privacy Tools To Improve Text Output Quality
• Understanding Aggregate Trends for Apple Intelligence Using Differential Privacy