Data security for AI
Research from Google suggests that with access to just 0.01% of a data set during training, attackers can poison an entire AI model. This risk escalates when multiple data sources are involved. How can organizations seamlessly navigate diverse platforms, maintain security and compliance, and protect data from breaches and manipulations—all without disrupting workflows or modifying code? Serge Vilvovsky will discuss the critical challenges of securing data collaboration and the AI supply chain.