Home/Responsible Artificial Intelligence Usage & Disclosure Policy

Responsible Artificial Intelligence Usage & Disclosure Policy

As an independent publication that prioritises engaging, first-person journalism and reporting, Man of Many has implemented strict policies regarding the use of Artificial Intelligence (AI). While AI tools are not explicitly banned within the Man of Many newsroom, the publication’s team of reporters, editors, video and social media producers are bound by the rules detailed in this policy.

As demonstrated across industries, the use of AI software may enable staff members to work fast, more efficiently and with greater creativity. These tools can often simplify menial or time-consuming tasks, providing enhanced efficiency for administrative purposes, concept ideation or story backgrounding. With technology and innovation core to the Man of Many business, having a firm understanding and actionable knowledge of ethical AI capabilities is critical to the ongoing development of the contemporary newsroom.

Man of Many acknowledges that modern journalism is evolving and embracing new technologies is pivotal, however, that does not negate the publication’s accountability. The publication remains committed to ensuring that all written and visual content meets the Man of Many standards of editorial ethics, integrity and transparency. As such, all Man of Many employees are subject to the following AI rules:

Reporting & Written Content

Man of Many believes that accurate, realistic and human-led reporting is critical to the ongoing success of the wider publishing industry and a vital resource in the fight against misinformation. No AI-written journalism is published on Man of Many. All reporting is original work by our journalists, published under their own byline. Some commercial and affiliate content is drafted using AI tools and then verified and released by a Man of Many editor, and it is published under the Man of Many byline rather than a journalist’s.

Journalists may use AI tools for research, transcription, formatting, structuring and ideation. They may also use them to compile factual reference material inside an article, such as a specification table, a comparison table or a product listing, which our team then checks against the original source. The reporting, the judgements and the article text are written by the journalist. A human editor proofreads and fact checks every piece before publication.

Keeping Articles Up to Date

Some of our articles are designed to stay current. Sales roundups, deals pages and annual buying guides are updated repeatedly after they are first published.

We use AI tools to keep the factual parts of those articles current: prices, availability, product links, deal details and formatting. The article keeps its original web address and its original author, because the writing is still theirs. What changes is the data, not the journalism.

If we ever rewrite the substance of an article using AI rather than updating its details, we treat that as new AI-drafted content and it is published under the Man of Many byline, not the original journalist’s.

Images

Man of Many will avoid publishing life-like AI-generated images on-site unless explicitly disclosed. AI-generative tools may be used for the purposes of image resizing, alt-tag naming conventions and curation.

Where a life-like AI-generated image is published, it carries a clear, reader-visible label in the caption (for example, “Image generated using AI” or “AI illustration”), not only in the alt text or file metadata. The label is legible to the reader on the page and is included in the caption so screen readers announce it.

Man of Many does not publish AI-generated or AI-altered images that depict real, identifiable people, including public figures, talent, staff or members of the public, without that person’s consent. The only exception is an image that is unmistakably satirical or clearly labelled as a non-photographic illustration and signed off by an editor. AI is never used to fabricate a photo-real depiction of a real person or a real event in a way a reader could mistake for genuine photography.

These rules apply to all reader-facing imagery, including social and newsletter assets.

In advertising and sponsored content we follow the advertising industry standard: any image or video generated from an AI prompt carries a label, whatever human editing follows, unless it is obviously not meant to look real. Routine editing that would not need disclosure if done by hand, such as colour correction, lighting adjustment, background cleanup or resolution improvement, is not separately labelled.

Disclosure

Man of Many understands and acknowledges the role of AI tools in the modern newsroom is developing rapidly. Under the stipulated guidelines, Man of Many journalists are permitted to utilise AI tools for the purposes of ideation, administrative tasks and supporting journalistic functions, however, are not to publish written AI content, unless the piece centres around the use of AI content. In this instance, the article will be clearly disclosed, and any errors will be clearly flagged. Importantly, all content, written and visual, will be subject to verification and approval by a human editor, ensuring the utmost accuracy, originality and transparency.

AI-Assisted Workflows

Man of Many acknowledges that the AI landscape has evolved significantly, with large language models (LLMs), AI agents and multimodal AI tools now widely used across the media industry. The publication uses AI-assisted tools for operational workflows such as data analysis, scheduling, content distribution and internal reporting. These operational uses do not affect editorial content and are subject to human oversight at all times.

Where AI tools are used in any reader-facing capacity, this is disclosed transparently and in accordance with the guidelines above.

Machine-Readable Marks and Provenance Signals

AI tool providers have begun embedding machine-readable marks in the content their models produce. From August 2026, text generated by some models carries an invisible watermark, and generated image files may carry signed provenance metadata. These marks are designed to survive copying and to be detectable by third parties.

Readers should understand what a mark of this kind does and does not tell them. It indicates that an AI tool was used somewhere in the process of making a piece of content. It does not identify who wrote that content, and the tool providers say so themselves: a mark can appear on work whose ideas, reporting and words came from a person, because AI tools are commonly used to proofread, translate, summarise or convert existing writing. Equally, the absence of a mark does not establish that a human wrote something.

Our answer to the authorship question is this policy. Man of Many journalists write Man of Many journalism, and it carries their name. Every piece is read, verified and released by a named human editor who is accountable for its accuracy and its voice. Where AI has drafted a significant element of what you see, we do not put a reporter’s byline on it.

If a reader, a regulator or a partner wants to understand how a specific Man of Many piece was made, we will answer the question directly.

The Standards We Have Drawn On

We wrote this policy ourselves, and it is our own position rather than a certification we have been granted. In shaping it we read the main industry standards on AI transparency and took from each the parts that fit an independent publisher:

  • The IAB AI Transparency & Disclosure Standards v2, the advertising industry’s framework for when AI involvement should be disclosed to consumers: iab.com/guidelines/ai-transparency-disclosure-standards-v2
  • The Coalition for Content Provenance and Authenticity (C2PA), the open standard for recording how digital content was created and edited: c2pa.org
  • The 100% Human initiative, which we have been a member of since 2024: 100human.com.au
  • ProudlyHuman, an independent certification scheme that publishes limits on acceptable AI use for creative work: proudlyhuman.org

These frameworks do not agree with each other on every point. Where they differ we have made our own call and explained it here.

Policy Review

This policy is reviewed annually and updated to reflect the evolving AI landscape, emerging best practices and regulatory developments. The most recent review was conducted in September 2026.

Read all of our Editorial Policies & Standards here.