AI Usage Disclosure
This page explains how ProductGuru uses artificial intelligence in producing content, what it does not use AI for, and how to tell us when something is wrong. Required and published in compliance with EU AI Act Article 50(4).
What AI does at ProductGuru
We use large language models (currently Groq Llama 3.3 70B for blog content scoring and Anthropic Claude as the operational assistant) to:
- Generate descriptive blog content from public-source data (government recall registries, regulatory ingredient lists, open product databases).
- Score and rank content quality so only the strongest gets published to Google's index.
- Translate non-English product titles to English.
- Help maintain the site (this is one of those things — yes, an AI helped publish this page).
What AI does not do at ProductGuru
Per our standing editorial rule, AI does not:
- Give medical advice. We surface facts about products, recalls, and ingredients; we always say "ask a pharmacist, doctor, or specialist."
- Give financial advice. We may surface product prices; we never tell anyone what to buy.
- Recommend by commission. Our editorial ranking is independent of our affiliate revenue.
- Fabricate facts. Every regulatory claim cites its government source with a last-fetched date and a deep link to the original.
Human editorial oversight
The jednatel of Czech Export Advisory s.r.o. (IČO 21928762) holds editorial responsibility for everything published on this site. Editorial decisions — what publishes, what doesn't, what gets corrected — are human, even when the drafting is AI-assisted.
If something is wrong
Models, providers, and where your data goes
The AI models we use run on third-party infrastructure (Groq Inc., Anthropic PBC — both US-based). Content generation processes public data (regulatory feeds, product catalogues). It does not process visitor personal data. Your searches, page views, and any future account data stay on our infrastructure in the European Union.
This page exists because the law requires it
EU AI Act Regulation (EU) 2024/1689, Article 50(4): deployers of AI systems that generate text published to inform the public on matters of public interest must disclose that the text has been artificially generated or manipulated. This page is that disclosure. Every page on this site carries a machine-readable <meta name="ai-generated"> tag and a visible footer notice on AI-assisted blog posts, both linking here.