AI Disclosure & Editorial Standards

CryptoKiller is transparent about how its investigations are produced. Our content is AI-drafted from collected evidence, passes automated source-verification gates, and is reviewed by a named human editor before it is published. AI accelerates how fast we turn verifiable evidence into a structured investigation — it never decides what is true, and every published page carries a named human reviewer who is accountable for it.

How our content is created

Our pipeline is evidence-first. We begin with evidence we have collected — captured advertisements, landing pages, domain and infrastructure records, regulator bulletins, and corroborated victim reports — and only then draft an investigation from that evidence. AI assists with drafting; it does not gather the facts and it is not permitted to invent them.

  • Evidence first — an investigation only begins once verifiable evidence has been collected and archived.
  • AI-assisted drafting — AI turns the collected evidence into a structured draft, with no invented facts, figures, or sources.
  • Automated quality and source-liveness gates — drafts pass automated checks that confirm the cited sources are real and reachable before a human sees them.
  • Human editorial review — a named human editor reviews the investigation against the evidence and is accountable for what is published.

Level of AI assistance

We classify our content as "AI-generated with human editorial review (L3 on the L0–L4 scale)." AI produces the draft from collected evidence, and a named human editor reviews it against that evidence before publication. The human reviewer — not the model — is accountable for the published page.

AI-generated images on this site carry the IPTC DigitalSourceType value trainedAlgorithmicMedia, identifying them as machine-generated. Screenshots of scam advertisements are captured evidence — they are real artefacts we recorded from live ad campaigns, not AI-generated images.

Accountability

John Feldt is our Editorial Standards Reviewer and is publicly accountable for our editorial standards. His professional profile is at https://www.linkedin.com/in/john-feldt-240838249/.

Individual analysts publish under consistent personas to protect their operational security, because scam operations are frequently run by organised groups that retaliate against investigators. Personas are stable and accountable — they are a security measure, not anonymity. The publisher of record for every investigation is DEX Algo Technologies Pte Ltd.

Corrections

If you believe we have published something inaccurate, email corrections@cryptokiller.org with the URL of the affected page and a clear explanation of the error. We review every correction request on the merits and publish a dated correction notice when warranted.

Back to home · Browse investigations · Read the blog