Cloudflare launches Identity-Aware AI Gateway to track who is using AI

Cloudflare has unveiled a new service called Identity-Aware AI Gateway that attaches a verified identity to every AI request leaving a company's network. This allows IT and security teams to see who specifically sent each AI prompt to a model provider. The service also enables setting spending limits per individual or team and automatically flags risky content. No additional systems are required to implement this.

Complementary to the AI Gateway is User Insights, which analyzes normal AI usage patterns for each user and system on a network. It then alerts teams when there are deviations from these patterns. With AI usage rapidly increasing and costs plummeting per request, companies face two main challenges - unpredictable expenses and lack of visibility into who is using AI systems. Cloudflare's solution addresses these issues by providing real-time visibility, budget controls, and security measures.

The company's existing zero-trust access product, Cloudflare Access, is integrated with the AI Gateway. This allows companies to drop shared API keys and authenticate AI requests against existing identities and network access setups. By funneling all AI traffic through a single point, redundant requests can be cached, rate limits can prevent runaway usage, and sensitive data like employee names and passwords can be stripped from the requests before they reach external AI providers.

The new AI Gateway is an extension of Cloudflare's existing security controls, with enhanced security features being added to the zero-trust platform since August of last year.