AI Disclosure Analysis

LLMSE analyzes websites for AI transparency compliance, detecting AI disclosure statements, AI usage policies, and chatbot labeling. Checks are conditional: issues are only raised when AI or chatbot technology is detected on the page. Sites with no AI footprint score 100/A.

AI Disclosure Grading

Every classified website with detected AI/chatbot technology receives an AI Disclosure grade:

A Excellent (90-100) — Full AI disclosure with explicit statements and policy links
B Good (80-89) — Partial disclosure present, minor gaps
C Fair (70-79) — AI technology detected with significant disclosure gaps
D Poor (60-69) — Multiple disclosure requirements missing
F Failing (<60) — AI/chatbot present with no disclosure or labeling

What We Detect

Three categories of AI disclosure signals are analyzed:

Visible Disclosure Text like “generated by AI”, “AI-assisted”, “powered by AI” in visible page content (scripts/styles excluded)
Policy Links Links to AI usage policy pages (e.g., /ai-policy, /ai-usage, /ai-transparency)
Chatbot Labeling AI attribution on chat widgets via aria-label, title, or visible text near widget containers

Scoring Logic

Deduction-based scoring starts at 100 points. Issues are only raised when AI/chatbot technology is detected:

Critical -15 points — Chatbot detected but not labeled as AI (conditional on chatbot tech present)
Warning -5 points — No visible AI disclosure statement found
Warning -5 points — No AI usage policy page link detected
Info -1 point — AI disclosure found but uses vague language

Full Reports via MCP & API

Get detailed AI disclosure reports with detected signals and recommendations:

MCP Server

Use the analyze_ai_disclosure tool through your AI assistant:

"Check AI disclosure for https://example.com"

Set up via the LLMSE Public MCP server.

REST API

Call the AI Disclosure endpoint directly:

GET /api/v1/ai-disclosure?url=https://example.com

See full parameters and response schema in the interactive API docs.

Browse by AI Disclosure Grade

Filter classified websites by their AI Disclosure grade:

ai_disclosure Grade A (Full Disclosure) ?ai_disclosure=A
ai_disclosure Grade B (Good) ?ai_disclosure=B
ai_disclosure Grade C (Fair) ?ai_disclosure=C
ai_disclosure Grade D (Poor) ?ai_disclosure=D
ai_disclosure Grade F (Failing) ?ai_disclosure=F

EU AI Act Context

The EU AI Act (effective August 2, 2026) requires transparency obligations for AI systems, including disclosure of AI-generated content and clear labeling of AI interactions. This analyzer helps assess readiness for these requirements by detecting existing disclosure practices on websites.

LLMSE detects 23 chatbot/AI technologies including Intercom, Drift, ChatGPT Widget, Dialogflow, and more. The language dimension enables natural filtering for EU markets (German, French, Spanish, Dutch, Italian, etc.).

Learn More