Vector Databases: AI Uses, Privacy Risks, and Mitigations | Privacy Engineering & Technology Education Discussion (PETed) Recording
Developers are rushing to adopt new AI tools and techniques, and private data for AI systems is shifting out of models and into vector databases. These new databases are immature from a security and privacy perspective, and the attacks against them are numerous and growing by the day. Understanding, controlling, monitoring, and protecting the data in these databases should be a top priority of security and privacy teams.