Corporations and governments increasingly use AI in ways that have profound impact on individuals. These systems are difficult to understand, even for those who create them.

One response to this situation is to call for a halt to AI development and to create regulatory guardrails around AI use. These are worthy discussions, but our belief is that when approached thoughtfully, AI can be used for regulation.

At webXray we use the same tools and techniques corporations use to track people, but instead of people, we track corporations. Our systems ingest huge volumes of information, make it searchable, and identify privacy sensitive content using our proprietary machine learning models.

Our approach vastly simplifies and accelerates compliance investigations.

For example, a simple keyword search for "cancer" may produce results related to research findings, astrology, fundraising, and more. This can slow down investigations as data protection law typically applies to information about specific individuals' treatment or conditions, rather than charity walk-a-thons.

The need for investigators to spend hundreds of hours to review results generated by keyword searches represents a significant barrier to finding privacy violations. We have experienced this problem in our own work and developed a solution.

Using webXray's powerful classification tools, investigators may combine a keyword search for "cancer", limit it to known hospital websites, and then apply our AI-powered "patient directed" filter. In this way it is easy to find potential violations of medical privacy laws such as HIPAA. This happens in seconds, not hours, and advances our mission of privacy acceleration.

Our current library of custom models, available to customers who upgrade →, flag content directed to children, patient-directed medical information, personal financial information, and spiritual beliefs. We are expanding our models to cover more protected classes of personal information, increasing their precision, and applying them at ever larger scale.

A final point about our AI approach is a focus on sustainability. We use clean energy to create very small, highly tuned, and efficient models. These models are then used to precompute privacy scores when we build our search index. This makes our systems faster, more accurate, and uses vastly less energy than other approaches.