Information Press is a California-focused news publication covering government, housing, transportation, business, finance, climate, healthcare, education, technology, consumer affairs, and community issues. This policy explains the limited role of artificial intelligence in the publication's visual work, the editorial activities for which we do not use AI, and the human responsibility for accuracy and publication.
Our approved approach is deliberately narrow: AI may be used to create visual illustrations and other image-based materials. AI is not used to write or draft articles, translate text, summarize documents, monitor sources, generate headlines, or make editorial decisions. Every published claim remains subject to source-based verification and human editorial oversight.
What AI Is Used For - and What It Is Not Used For
Illustrative, Not Documentary, Content
Information Press may use image-generation tools to create illustrative artwork for subjects that benefit from conceptual rather than documentary representation. This may include abstract graphics about consumer technology, housing affordability, financial concepts, business trends, education, or environmental themes. The distinction concerns the role of an illustration in a news story: an image may support explanation without serving as a source for the underlying facts. Readers should not have to guess whether a scene is documentary evidence.
An AI-generated visual should help clarify or illustrate a subject without presenting an invented scene as an actual event. Its use must be appropriate to the article, the likely understanding of readers, and the potential consequences of ambiguity. Where an actual photograph, official map, or sourced diagram is needed for factual reporting, a synthetic approximation is not a substitute.
Information Press does not use AI to generate completed news stories, explainers, analysis, guides, or consumer reviews. It does not use AI for first drafts, translations, summaries of source documents, automated source monitoring, editorial research organization, headline alternatives, grammar or style rewrites, or automated classification and tagging of editorial articles.
We also do not authorize automated publication by AI systems. Software may support ordinary website functions, feeds, publishing operations, or manual research without implying that AI produced the accompanying reporting. A change to these stated boundaries would require an updated public policy before being represented as an accepted practice.
Human Responsibility and the Ban on Fabricated Evidence
What Human Editorial Oversight Covers
Human editorial responsibility remains with Information Press for selecting subjects, establishing factual claims, evaluating evidence, determining an article's context, reviewing visuals, choosing labels, and approving publication. A generated illustration should receive a review appropriate to the subject and risk of misleading readers. The presence of AI-generated imagery does not lower the standard for factual statements elsewhere on the page. Names, quotations, statistics, incidents, and citations require an authentic evidentiary basis.
Review may include examining whether the image introduces nonexistent objects, misleading symbols, fake documents, incorrect geography, real-looking persons, an invented accident scene, or a representation that conflicts with the article. Illustrations should be rejected, revised, labeled, or replaced if they create a material false impression. Human oversight is not a claim that every mistake can be prevented.
What Cannot Be Invented
Information Press prohibits the use of AI to invent sources, quotations, interviews, research, public records, regulations, statistical findings, eyewitness accounts, organizations, awards, professional qualifications, or real events. AI-generated text must not be laundered into an article as if it were an independently checked source.
Synthetic photographs or visual artifacts must not be presented as evidence that something happened. An image of a purported police scene, wildfire, eviction notice, hospital, public demonstration, or government action does not establish the existence of that event. Editorial assertions require appropriate direct or documentary evidence apart from any illustration.
How AI-Generated Images Are Described and Displayed
Illustrations Versus Documentary Photography
AI-generated illustrations are conceptual or artistic material. They are not documentary photography. Readers should be able to distinguish a synthetic illustration from a photograph of an actual location, individual, disaster, public meeting, or news event when confusion is reasonably possible. A caption should avoid wording that places a synthetic scene at a real event or implies a person was photographed there. The label and surrounding context should be consistent with what the image actually represents.
For reporting on actual breaking events, public emergencies, elections, protests, injuries, or alleged misconduct, authentic verified source imagery should generally be preferred when available and appropriate. The absence of authentic imagery does not justify creating a realistic false visual record. Images must not imply access to a scene that the publication did not document.
Disclosure, Captions, and Alternative Text
Visible Labels and Accurate Descriptions
When an AI-generated image could reasonably be mistaken for a real photograph or materially affect interpretation, an accurate visible label should be included. Appropriate captions include “AI-Generated Illustration,” “Conceptual AI Illustration,” or “Illustrative Image Created with AI.” Captions must not claim that a synthetic depiction shows the actual event described in the story.
Image alternative text should describe the visible subject honestly and, when necessary, identify that it is synthetic. An illustration credit or label should not invent a photographer, design agency, license, or source. The choice of label should reflect the degree of realism and the risk of a misunderstanding, not merely aesthetic preference.
Real People, Institutions, and Impersonation
AI illustrations must not falsely depict an identifiable real person performing an action, making a statement, endorsing a product, participating in a protest, committing misconduct, or appearing at a location where they were not documented. Synthetic images must not impersonate government officials, public agencies, doctors, financial professionals, or other people in ways that create false authority.
Visuals should not fabricate official identification, medical scans, bank statements, court filings, ballots, evacuation maps, or institutional records. Where an illustration uses general symbolic elements, editors should assess whether its context or layout might still suggest a genuine document or evidence.
Accuracy, Rights, and Supporting Sources
Visual Accuracy and Use Rights
AI-generated visuals may contain distortions, false text, inaccurate emblems, unlikely objects, or mistaken representations of cultural and geographic details. These risks should be examined during editorial review, especially where they affect a factual interpretation. Cropping, captioning, and placement can also create misleading implications even if the underlying image is presented as an illustration. A visually plausible result is not confirmation that a depicted object, location, person, or sequence of events existed. Source checks must concern the underlying claim, not the apparent realism of the image.
The use of an image-generation service does not automatically establish exclusive copyright ownership, an unrestricted commercial license, or permission to imitate a protected work or identifiable individual. Image use should be evaluated against applicable service terms, intellectual-property rights, and relevant publication practices. We do not claim a proprietary licensed archive or external image partnership without documentation.
Source Verification Is Separate from Visual Production
Articles are evaluated according to underlying evidence rather than the apparent plausibility of a generated image. An illustration does not replace primary sources, official reports, research, attributable expert information, or factual checking. Editorial claims involving dates, figures, quotations, jurisdiction, or identified people require their own support.
Information Press prioritizes named and verifiable sources and rarely considers confidential sources. Anonymous or synthetic identities must never be created to give a story invented corroboration. Our Fact-Checking Policy and Unnamed Sources Policy describe the applicable standards.
Corrections, Questions, and Policy Updates
Reporting Misleading Visuals
If an illustration contains a material factual inconsistency, inadequate disclosure, misleading contextual implication, or an inaccurate caption, Information Press may relabel, correct, replace, or remove it as appropriate. Significant issues may warrant a visible editorial explanation, in line with the Corrections Policy .
Readers can report concerns to corrections@informationpress.net or contact the editorial address earthday@informationpress.net . Include the article URL, visual or caption in question, the specific concern, and relevant supporting information. A concern will be assessed for its factual significance rather than dismissed merely because an image is labeled illustrative.
Related Editorial Standards and Future Changes
Information Press may update this policy if its approved uses of AI, editorial technology, or disclosure standards change. An expanded use should be accurately described at the time it becomes operational rather than implied retroactively. This policy does not claim that generated illustrations have already appeared in any particular article.
- Editorial Policy
- Publishing Principles
- Ethics Policy
- Fact-Checking Policy
- Corrections Policy
- Contact
- Diversity Policy
- Privacy Policy
The publisher, Information Press, remains accountable for visual accuracy and editorial decisions at https://informationpress.net/ .