AI Disclosure
Last updated
AI Disclosure
We use AI as a research and drafting tool. We do not use AI as a substitute for editorial judgement.
What AI does for us
- Discovery. Continuous scraping of regulator press releases, exchange filings, conference RSS feeds, and authoritative trade press. AI clusters duplicate stories and surfaces candidates for editorial review.
- Triage and scoring. A scoring model rates candidate stories on novelty, regional relevance, search demand, and citation potential. The bottom 70% never reach an editor.
- Drafting. AI produces a structured first draft from verified sources. Every numerical claim must trace back to a scraped source URL.
- Fact-check pass. A second AI agent re-reads the draft, extracts every claim, and verifies it against cited sources. Discrepancies are flagged for the human editor.
What only humans do
- Editorial review. Every article is reviewed and edited by a named human within 24 hours of agent draft. The editor adds 2–3 sentences of original perspective — a stat from our database, a connection to another story, an industry contact's reaction — that meaningfully improves the article.
- Final fact-check on sensitive claims. Quotes, attributions, named individuals, and any claim that could cause reputational harm are verified by the human editor against the original source.
- Publication decision. No article is published without explicit human sign-off.
- Right of reply. Outreach to subjects of critical coverage is done by humans.
What we never do
- We do not publish unedited AI output.
- We do not fabricate quotes or attributions. Quotes are extracted from sources that the agent has scraped, with attribution and source URL preserved.
- We do not "rewrite" competitor articles. Our coverage starts from primary sources.
- We do not hide our use of AI. Every article carries a footer disclosing AI-assisted research with named human review.
Disclosure footer
Every article ends with:
Researched and drafted with AI assistance; reviewed and edited by the named author within 24 hours of draft. See our editorial standards and AI disclosure.
If you find an article that does not carry this footer or that you suspect was published in error, please email earthenergylog@gmail.com.
Why we work this way
Google's spam policies and the March 2026 Core Update penalise "scaled content abuse" — sites publishing AI articles at volume without editorial oversight. The publications that survived 2025–2026 algorithm updates are the ones that use AI for research and drafting but ship only content that has cleared human editorial review.
We believe this is also the right way to publish. AI without human judgement produces output that is statistically plausible but often subtly wrong. A named editor adds accountability, context, and the original perspective that turns an article from a summary of public information into something worth reading.