EU AI Act Website Scanner
NotSoHuman.ai's EU AI Act Website Scanner is a paid web-based service that crawls a public website and reports on AI-related content and features that may fall under the EU AI Act's Article 50 transparency obligations. Given a URL, the scanner reviews pages for AI-written text, AI-generated or manipulated imagery (including deepfakes), AI chatbots, and embedded AI widgets or plugins, then flags items that may need labelling, disclosure, or further human review. The product is positioned around a straightforward three-step workflow: enter a website URL, review the generated report of AI findings and transparency gaps, then apply the guidance on labels and disclosures. Reports are framed as indicative AI transparency signals intended for internal review rather than as legal advice, and the vendor says the tool is used by more than 3,000 businesses. The homepage states the tool is aimed at website owners, digital and agency teams managing multiple client sites, and legal or compliance reviewers who need a focused starting list of pages and features requiring human judgement. Agency use is explicitly supported through a workflow for checking several client websites. NotSoHuman is delivered entirely as a hosted SaaS product accessed through a browser login, with separate marketing pages for the product, EU AI Act guidance, pricing, and a blog. No public API, on-premise deployment, or open-source repository is described in the available homepage material.
- Category
- AI/ML
- Pricing
- Paid
Problem solved
AI-generated text, imagery, chatbots, and third-party AI widgets are increasingly embedded across public websites, making it difficult for organisations to track where EU AI Act Article 50 transparency disclosures may be required. Manually auditing large sites for these signals is slow and error-prone. The scanner automates discovery of AI content and AI features and converts findings into a reviewable list of items that may need labelling or disclosure.
Use cases
- Audit a company website for AI-written text that may require transparency disclosure under EU AI Act Article 50
- Detect AI-generated or manipulated images and deepfakes published on marketing pages
- Inventory AI chatbots, assistants, and embedded AI widgets across a site
- Run compliance-oriented website reviews across multiple client sites from an agency workflow
- Produce a page-level list of AI findings for legal or compliance teams to review manually
- Pre-launch check of new site sections or campaign pages for unlabelled AI content
Audience
- Business Users
- Enterprise
- Product Managers
- Content Creators
Pros
- Purpose-built around EU AI Act Article 50 transparency rather than generic AI detection
- Covers multiple signal types in one pass: AI text, AI images, deepfakes, chatbots, and AI widgets
- Report includes recommended actions (label, disclose, or review further) instead of raw scores only
- Agency-oriented workflow allows checking multiple client websites from one account
- No installation required; scanning starts from a URL in the browser
Cons
- Vendor states output is indicative and not legal advice, so expert compliance review is still required
- Paid subscription product with no free or open-source tier described on the homepage
- Detection of AI-generated content is inherently probabilistic and can produce false positives or misses
- Scope is limited to EU AI Act Article 50 transparency signals rather than broader AI governance or other regulations
- No public API, on-premise option, or repository is documented in the available homepage material
Official website: https://notsohuman.io/
Canonical LambdaBase page: https://www.lambdabase.com/tools/eu-ai-act-website-scanner