How Genius Eye News Uses Automation

Draft for attorney review. This document has not been approved by a licensed attorney and is not yet in force. Items marked [CONFIRM] describe intended product behavior that engineering and editors must verify before launch.

Effective date: [DATE OF LAUNCH] Last updated: [DATE OF LAUNCH]

Genius Eye News, operated by [LEGAL ENTITY NAME] (d/b/a TRG Tech Link), uses automation — including artificial intelligence (AI) — to report the news. We believe readers deserve to know exactly how. This page explains what our software does, what people do, how we label automated work, and the lines we will not cross.

1. The short version

Most of our stories are drafted by automated newsroom software ("newsroom agents"). The software is not a person, and we never present it as one.
Our sources are public: official government data and other news outlets' public reporting. We always link our sources, and we never copy other outlets' text or photos.
Every story passes an automated fact-check step before publication.
Human editors oversee the system, review flagged and sensitive stories, and can correct or unpublish anything. Corrections are public.
Digital anchors and synthetic voices are labeled as AI-generated. We never create a synthetic version of a real person.

2. What our newsroom agents do

Step	What the software does
Monitor	Watches official public data feeds — for example the National Weather Service (NWS), National Hurricane Center (NHC), U.S. Geological Survey (USGS), FEMA, the Securities and Exchange Commission (SEC/EDGAR), and the Federal Register — and reads public reporting from other news outlets
Draft	Writes an original story in our own words, summarizing the facts and attributing them ("according to the National Weather Service…," "[Outlet] reported…") with links to every source
Fact-check	A separate automated step checks each factual claim in the draft against the cited sources — names, numbers, dates, places, and quotes — and checks that attributions are accurate. Drafts that fail are blocked or sent to an editor
Classify	Tags the story by city edition and section, and flags sensitive topics for human review
Publish and update	Publishes stories that pass, and updates developing stories as official data changes

What our software does not do: It does not interview people, attend events, or claim to have done so. It does not invent quotes; any quote in a story is attributed to the outlet or official document where it appeared, with a link. It does not reuse other outlets' photos — our images are official public-domain government imagery (attributed), licensed images, embedded official live streams, or clearly labeled illustrations.

3. What people do

Editors set the rules the software follows, monitor what it publishes, and can correct or unpublish any story at any time.
Human review before publication is required for stories the system flags as sensitive, including [CONFIRM list]: crime stories that name a private individual; allegations of wrongdoing against a named person or business; deaths and serious injuries; stories involving minors; election results and election procedures; and any story about a Genius Eye affiliated company.
Human review after publication applies to reader correction requests, which are always decided by an editor (see our Corrections Policy).
Advertising and classifieds are screened by software, but political ads, financial-services ads, and sponsored articles are always reviewed by a person, and anyone can request human review of an automated rejection.

4. How we label automated work

Bylines. Stories drafted by our newsroom agents carry the byline "Genius Eye News Automated Desk" with a link to this page, and the footer note: "This story was drafted by automated newsroom software from the sources linked above, checked by an automated fact-check step, and is subject to editor review. Report an error." [CONFIRM byline wording.]
Human-edited stories. If an editor substantially rewrites or adds reporting to a story, the byline reflects that ("Automated Desk, with editing by [Name]").
World Watch. World Watch scores and "what it could mean" analysis are generated by an automated model and labeled "Automated analysis — not investment advice."
Market data and summaries. Automated market summaries are labeled as automated, and market data is labeled as delayed.
Translations (if offered) are labeled as machine translations.
Images. AI-generated illustrations are labeled "AI-generated illustration" in the caption and are never used to depict a real news event as if it were a photograph.

5. Digital anchors and synthetic voices

Some of our audio and video segments are presented by digital anchors — computer-generated presenters — or read by synthetic (AI) voices.

Always labeled. Every segment with a digital anchor or synthetic voice displays an on-screen label such as "AI-generated presenter" for its full duration (or at the start and end of audio-only segments, plus a spoken disclosure: "This report is read by an AI voice."). The segment's page repeats the disclosure.
Fictional identities only. Our digital anchors are original characters. They are not modeled on, and do not imitate, the face, voice, or name of any real person — living or dead — including journalists, celebrities, or public officials.
No synthetic real people, ever. We do not create AI-generated or manipulated images, audio, or video of real people saying or doing things they did not say or do. When news requires showing a real person, we use authentic, attributed material (for example, an official live stream or public-domain government footage).
Synthetic voices of real people are prohibited, even with consent, in news content. [Business decision — confirm.]
No digital anchors in ads. Digital anchors do not endorse products. If a digital anchor reads a sponsor message, it is labeled "Sponsor message" on screen, and digital anchors never appear in political ads.
Content credentials. Where technically feasible, AI-generated images, audio, and video we publish carry embedded provenance metadata (such as C2PA content credentials) identifying them as AI-generated. [CONFIRM.]

6. Third-party live streams and camera images

Live pages embed official third-party live streams (for example, government or agency YouTube streams) and government traffic-camera images. We attribute each one to its source, we do not alter them, and we do not control their content or availability.

7. What our AI does not use

Our newsroom software works from public information. It does not use readers' personal information, newsletter lists, or classified listings to write stories or to train AI models. [CONFIRM — including AI vendor contract terms.]
Content you send us (for example, a news tip) may inform a story, but is reviewed by an editor before use.

8. Accuracy, limits, and accountability

Automation lets us cover more communities quickly, but it can make mistakes — misreading a data feed, mis-summarizing another outlet's report, or missing context. That is why we link every source, run a fact-check step, keep editors in control, label automated work, and publish every correction. Genius Eye News — not the software — is responsible for what we publish. If you see a mistake, please report it.

9. Laws and standards we follow

We design these practices to be consistent with the FTC Act's prohibition on deceptive practices (15 U.S.C. § 45), state laws on AI disclosure and synthetic media (for example, the California AI Transparency Act, Cal. Bus. & Prof. Code § 22757 et seq., and state election deepfake laws described in our Political Advertising Policy), and state right-of-publicity laws. We review this page as laws and our technology change.

10. Questions

news@geniuseyenews.com · [LEGAL ENTITY NAME], [MAILING ADDRESS]