AI policy
Last updated: 07/05/2026
Robots Atlas uses large language models (LLMs) as tools in the content creation process. We believe in full transparency with our readers — this document describes exactly how and where we use AI.
How we use AI
Research
AI helps us analyze and summarize large amounts of source material (scientific papers, documentation, company reports). It shortens the time needed to understand the context of a new topic.
First drafts
First drafts of articles are often generated with LLM assistance based on the collected material. The editor then verifies, edits and supplements the content.
Translations PL ↔ EN
We translate articles between Polish (robocikowo.com) and English (robotsatlas.com) with AI assistance. English versions are reviewed by a dedicated editor — an MA in English philology — for language naturalness and precision of technical terms.
Editorial oversight
Every publication is verified by an editor before release. We check facts, sources, dates, numbers and technical parameters. The editor takes full responsibility for the publication — we do not publish AI-generated content without human oversight.
Editorial process
Every article on Robots Atlas goes through five steps before it goes live:
1. Topic and source selection
An editor picks the topic from publications of the last 24 hours (company announcements, scientific papers, technical documentation). Every article must have at least one cited, verifiable source — without it, the article is not written.
2. AI-assisted research
Language models help us summarize and compare source material. AI is not a source of facts — it is a tool for working faster with sources we already have.
3. First draft
The article draft is produced with LLM assistance based on the collected sources. Every claim in the text must be traceable to a specific entry in the source list.
4. Human editorial review
An editor reads the full text and compares claims against the original sources. We check facts, dates, numbers, technical parameters, model names and company names. AI hallucinations are removed — when in doubt, the passage is cut or rewritten.
5. Publication and corrections
After publication we track reader reports. Every correction is documented in the Corrections section with a date and a description of the change — we do not hide changes retroactively.
Editorial pen names
Articles on Robots Atlas are published under pen names: Pan Robocik (AI topics) and Pani Robocik (robotics topics). Behind each pen name is a real person — an editor who prefers to keep their privacy. AI is a working tool, but every article is written and reviewed by a human.
Limitations
Language models can hallucinate — generate information that sounds credible but is factually incorrect. Our verification process is meant to catch this, but errors happen. If you spot an inaccuracy, please report it: Report a correction.
What models we use
Currently we mainly use models from the Claude family (Anthropic). The model list is updated as the ecosystem evolves.
Contact
Questions about our AI policy? Write to us: editorial@robotsatlas.com