AI Transparency & Editorial Policy

Last updated: May 2026

Raw Science operates on three principles: original analysis, verified citations, and transparent methodology. AI tools are part of my workflow, and this policy documents their roles and boundaries.

Where I use AI

Visual content. Header images are AI-generated for illustrative purposes only. However, every chart, diagram, and technical figure inside an article is either produced by me from primary data (using either Python, OriginLab, Excel, or Affinity software) or sourced from a cited reference.

News monitoring. I run an AI-assisted pipeline, built on a commercial API, that parses global news in several languages on the topics Raw Science covers, i.e., critical minerals, energy systems, materials, and supply chains. It filters items by relevance and language. Every item that appears in an article or newsletter has been read personally by me in the original source.

Language and grammar. Drafts may be polished for grammar, clarity, or English-Spanish flow using AI assistance. Arguments, structure, and substantive claims remain mine.

Where I do not use AI

Analysis and interpretation. AI does not generate arguments, claims, or conclusions. I want to draw particular attention to the Strategic Implications sections, the analytical core of the deep-dives: these sections are written by me, drawing on a decade of research across electrochemistry, materials science, photocatalysis, and corrosion science. This background drives the technical understanding, source evaluation, and scientific rigor behind every conclusion.

Citations. Every numbered reference corresponds to a source I have read in full. AI is not used to generate, verify, or summarize citations.

Accountability

The analysis you read in Raw Science reflects human judgment developed through academic training and primary source engagement. It will inevitably sometimes be right and sometimes be wrong, but it will be grounded in sources I have personally reviewed and remain fully open to scrutiny, challenge, and correction.

If you ever spot a claim that does not hold up, please tell me. That feedback loop is what makes this work.

— Giada