Foundation of AI Model Training
AI models can quickly learn users’ behavioral patterns, preferences, and language habits to make services more ‘personalized.’ Social media platforms provide user-targeted advertisements and content, and voice assistants can suggest convenient services tailored to user characteristics.
Background of the Problem
Users’ personal information is collected and stored on a massive scale to provide such personalized services. There is growing anxiety that online records, location information, and voice/facial recognition data could be used for AI training without the knowledge of the owners of the data. Moreover, once collected, personal information often isn’t completely deleted, leading to the risks of leaks and privacy violations.
