FingerprintJS unveiled previews of 'AI Assistant Detection' and the 'Automation Intelligence API' to filter traffic from major AI assistants such as ChatGPT, Gemini and Anthropic's Claude.

SiliconANGLE reported on June 1 that the two products focus on closing detection gaps created by a rise in AI requests that do not go through a browser.

'AI Assistant Detection' identifies major AI assistant web traffic in real time. It runs on the 'Automation Intelligence API', which identifies automated traffic without client-side JavaScript.

FingerprintJS released AI Agent Detection in February to identify autonomous AI agents that act on behalf of users. The newly unveiled tools expand detection to AI assistants that browse the web and summarise content. Security teams can view both types of AI-based access on a single screen.

Existing web security and analytics tools were designed on the premise that people connect by opening a browser. AI assistants, however, access websites through HTTP requests without loading pages, and fetch and summarise content or perform research tasks. Gemini Spark, which Google announced at I/O 2026, also operates on a dedicated cloud virtual machine without users opening a browser.

SiliconANGLE said scrapers and low-quality bots that spoof popular AI assistant user agents to bypass existing bot defences are increasing by exploiting this gap.

'AI Assistant Detection' works at the HTTP level instead of relying on browser signals. Companies can see which AI tools access their content and filter them out before they cause data scraping or analytics distortion. It verifies ChatGPT and Claude by matching IP ranges and verifies Gemini using a DNS-based method.

The 'Automation Intelligence API' can be deployed at content delivery network edges and middleware and across cloud backends. Along with classification results, it also provides IP and network risk signals such as proxies, virtual private networks, Tor and location information. Companies can use this to choose responses such as rate limiting, additional verification or allowing access instead of a blanket block.

FingerprintJS has raised a total of $77 million to date, including $32 million in funding secured in November 2021.