An AI-operated publication
AIViewer’s research, drafting, editing, illustration, coding, and maintenance may be performed substantially by AI agents. The founder sets the mission and owns the publication, but readers should not assume that a human editor has reviewed every sentence before publication. We state the production mode on each new editorial page instead of hiding it behind a generic “editorial team” byline.
Autonomy does not lower the evidence standard. “The model produced it” is never an excuse for an invented source, copied passage, unsupported claim, fake test, or misleading recommendation. The public page must show enough source and limitation information for a reader to challenge the work.
Acceptable uses
Editors and developers may use AI tools to assist with work such as:
- organizing notes, transcripts, or source material;
- suggesting outlines, questions, counterarguments, or plain-language explanations;
- producing a source-backed draft and checking its citations, structure, internal links, and policy requirements;
- copyediting, translation support, metadata, and accessibility text;
- extracting structured fields from verified sources, subject to checking;
- supporting code, data analysis, design exploration, and quality assurance; and
- creating clearly contextualized illustrations or other visual assets.
Uses we do not permit
- Publishing unverified AI output as fact or disguising materially AI-assisted work as solely human-authored.
- Inventing or guessing facts, statistics, quotations, entities, dates, URLs, test results, or credentials.
- Presenting generated commentary as first-hand product testing.
- Assigning a review score solely because an automated system inferred one from metadata or marketing claims.
- Rewriting another publisher's work to disguise its origin.
- Using automation primarily to create large numbers of search-targeted pages without distinct reader value.
- Creating a fake author, source, testimonial, screenshot, or user experience.
Page-by-page verification requirements
AI output is never treated as a source. Consequential factual claims are checked against current primary evidence where it exists. A source link must lead to the material it is cited for. Product conclusions must match the access and testing actually performed. When evidence remains uncertain, the page says so or removes the claim. Automated checks can detect missing fields, stale review dates, broken links, and duplicate routes, but they cannot turn weak evidence into a fact.
Disclosure is the default
New lessons, signals, foundations, and workflows include a page-level AI-use statement. It identifies whether AI materially performed research, source review, drafting, editing, translation, or image generation. We do not use “human-reviewed,” “expert-reviewed,” or a named reviewer unless that review actually occurred.
A named human may occasionally contribute or review a page. That is additional provenance, not a standing claim about the site.
Editorial covers and the Imagin workflow
AIViewer may use Imagin, our prompt-and-style workflow, to develop original directions for AI-generated editorial cover images. Imagin helps construct the scene, composition, mood, cultural details, and safety constraints; a separate image-generation model renders the final image. We do not describe Imagin itself as the system that generates the pixels.
These covers are illustrations. They are not evidence, documentary photographs, records of a real event, or proof that AIViewer tested a product. A depicted person, place, object, or moment should not be understood as real unless the page explicitly identifies and sources an authentic image.
Originality and factual boundaries for generated visuals
- We create a new editorial concept rather than copying a creator’s video thumbnail, composition, branding, or distinctive visual treatment.
- Generated covers must not contain logos, visible text, watermarks, fake quotations, or other marks that could imply a company, creator, or person supplied or endorsed the image.
- We do not use a generated cover as a fabricated screenshot, product interface, dashboard, chart, dataset, test artifact, or demonstration of what a tool produced.
- Authentic product screenshots may be cropped or redacted for privacy, but not altered to misrepresent what the product did.
People, culture, rights, and safety
Imagin’s visual direction is culturally grounded in ordinary Pakistani and broader South Asian settings, clothing, gestures, and light. Cultural specificity should feel plausible and respectful, not exoticized, flattened into a generic “South Asian” look, or built from religious, class, gender, regional, or ethnic stereotypes.
We do not intentionally place an identifiable private person in an invented scene or generate a recognizable public figure in a fictional or unverifiable scene. A generated likeness must never imply that a person performed an action, attended an event, expressed a view, endorsed AIViewer, or endorsed a product. When identity is necessary to the reporting, we prefer an authentic licensed image, an official embed, or a clearly non-identifying conceptual illustration. Images receive rights, privacy, impersonation, dignity, bias, and safety checks before use, with additional care for children, sensitive traits, traumatic events, and sexual or violent content.
Disclosure, accessibility, and creator attribution
We label an image as AI-generated or illustrative where a reasonable reader could mistake it for a real photograph, event, person, product interface, or piece of evidence. Page-level AI-use disclosures also identify material image generation. Alt text describes the illustration and its relevant visual meaning without presenting the scene as factual—for example, “AI-generated illustration of…” rather than wording that implies a documented person or event.
A generated cover does not replace source attribution. When a page embeds or discusses a creator’s video, the creator and original video remain credited separately, and the surrounding illustration is not presented as the creator’s thumbnail or work.
Image provenance, corrections, and removal
For generated editorial covers, we retain an internal provenance record: the page or asset identifier, prompt or prompt summary, creation date, image-generation model or service, material edits, and disclosure status. These records help us audit originality, reproduce the production history, and investigate concerns; they are not evidence that the depicted scene occurred.
Readers and rights holders may report a misleading image, disputed likeness, cultural concern, missing disclosure, attribution problem, accessibility issue, or possible rights violation through our Corrections Policy. We may relabel, replace, correct, de-index, or remove the image or page depending on the evidence, risk, and severity.
Structured and automated pages
Automation may organize verified facts into a consistent reference format. A structured reference page is not automatically a review, recommendation, or proof of first-hand experience. It should be indexed or monetized only when it offers enough original commentary, curation, analysis, or practical utility to stand on its own for readers.
Privacy and confidential information
Editors should not place confidential business material, unpublished reader correspondence, account credentials, or unnecessary personal information into third-party AI systems. When a legitimate workflow requires sensitive material, it should be minimized or de-identified and handled under the applicable provider terms and our Privacy Policy.
Authorship, copyright, and attribution
AI assistance does not make copied expression original. AIViewer attributes creators and external reporting, respects copyright and licenses, avoids close imitation of a living creator’s distinctive work, and uses official embeds rather than republishing transcripts or copied thumbnails. The publisher is identified as AIViewer; a named human contribution is credited only when it occurred.
Legacy content and enforcement
This policy applies to content published or materially updated on or after July 25, 2026. Older pages are being consolidated or removed from the public index. If we find fabricated sourcing, false testing claims, copied expression, or content that cannot stand on its evidence, we will correct, relabel, de-index, rewrite, or remove it.
Ask about our process
If a page leaves you unsure how AI was used, contact hi@aiviewer.ai with the URL and your question.