Bharatpedia:Artificial intelligence

"BP:AI", "BP:AICOMP", "BP:LLM", and "BP:CHATGPT" redirect here. For other uses, see BP:AI (disambiguation), BP:AICOMP (disambiguation), BP:LLM (disambiguation), and BP:CHATGPT (disambiguation).

Important note:

 AI is a tool, not an author.

You may use artificial intelligence to assist with routine editing tasks, but every edit you publish remains your own. You are fully responsible for anything you submit, whether or not AI helped produce it.

 Never blame AI for errors.

Claiming that a factual error, copyright violation, or policy breach "came from AI" is not a defence. Sanctions apply to the human editor, not to the tool.

 No AI-generated media in mainspace.

AI must not be used to generate images or video of real people or events. All such content is prohibited.

 Disclose when required.

Substantial AI assistance beyond grammar, spelling and formatting must be disclosed on the article's talk page or in the edit summary.

This page documents Bharatpedia's policy on the use of artificial intelligence (AI), including large language models (LLMs), generative text tools, AI image and media generators, machine translation systems, and similar technologies. It applies to all editors, administrators, functionaries, bots, and other contributors across all namespaces. Bharatpedia recognises that AI tools can meaningfully assist human editors with routine tasks. At the same time, AI systems can produce confident but incorrect output, fabricate sources, reproduce copyrighted material, and amplify bias. This policy therefore aims to permit beneficial uses of AI while holding human editors accountable for every edit they publish.

Purpose and scope

This policy applies to:

• text generation and editing, including drafting, paraphrasing, copy-editing, and translation;
• image, audio, video, and other media generation or manipulation;
• research assistance, including summarising sources and identifying citations;
• detection and maintenance tools, including anti-vandalism, reference checking, and categorisation;
• bots, gadgets, and scripts that incorporate AI models;
• administrative and discussion processes, including deletion discussions, requests for comment, and arbitration.

This policy does not restrict the use of conventional editing tools that do not rely on generative AI. It also does not restrict editors from discussing AI, criticising its use, or proposing changes to this policy through the normal consensus process.

Guiding principles

Four principles guide the interpretation of this policy:

1. Human responsibility. Every edit on Bharatpedia is made by a human editor. AI may assist, but responsibility for content, accuracy, licensing, and conduct rests entirely with the editor who publishes the edit.
2. Transparency. Where AI has materially shaped content, editors should disclose that fact so that others can review the output critically.
3. Verifiability. All content must remain verifiable through reliable, published sources. AI-generated text that cannot be traced to a real source is unacceptable.
4. Proportionality. The more consequential the edit, the greater the care required. A spelling fix needs no ceremony; a new article requires full human review.

Permitted uses of AI

AI tools may be used to assist with the following tasks, provided that the editor reviews and takes responsibility for the output.

Copy-editing and language

• Correcting spelling, grammar, punctuation, and capitalisation.
• Improving sentence clarity and readability without changing meaning.
• Converting between English varieties (for example, Indian English and American English) consistently within an article.
• Fixing awkward phrasing introduced by translation or by the editor's own writing.
• Suggesting alternative word choices where the original is ambiguous.

Formatting and maintenance

• Converting plain text into wiki markup.
• Fixing reference syntax, citation templates, and parameter names.
• Detecting duplicate references and consolidating them.
• Adding or correcting category tags, hatnotes, and maintenance templates.
• Cleaning up tables, infoboxes, and list formatting.
• Fixing broken internal links, section anchors, and redirects.

Grouping and organisation

• Grouping references, categories, or list items into a logical order.
• Reorganising article sections without altering content.
• Sorting table rows, bibliography entries, or external links.
• Identifying duplicate or overlapping sections for human review.

Research assistance

• Suggesting search terms for finding reliable sources.
• Summarising a source that the editor has already read, for the editor's own understanding.
• Translating a foreign-language source for the editor's own reading, provided the editor verifies key claims.
• Identifying likely relevant policies and guidelines for a given dispute.

Translation assistance

• Producing a first-draft translation from another language, which the editor must then verify line by line against the source.
• Suggesting idiomatic equivalents for technical terms.
• Checking that a translation does not accidentally introduce claims absent from the original.

Technical and administrative tasks

• Drafting bot code, gadget scripts, or template logic that a human reviews before deployment.
• Generating test cases, edge-case examples, or documentation for templates and modules.
• Assisting with routine maintenance reports, cleanup listings, and tracking categories.
• Summarising long discussions for the editor's own understanding — but not for posting as a close rationale without independent review.

Prohibited uses of AI

The following uses are not permitted under any circumstance.

Fabrication of content

• Generating article prose and publishing it without verifying every claim against real sources.
• Inventing sources, quotations, page numbers, DOIs, ISBNs, or publication details.
• Creating fake citations, fake interviews, or fake attributions.
• Producing biographical content about real people based on AI-generated inference rather than verified sources.
• Generating content about living persons without careful human review and sourcing.

Copyright and licensing

• Submitting AI output that reproduces substantial portions of copyrighted text.
• Uploading AI-generated images that imitate a living artist's style in a manner that infringes their rights.
• Uploading AI-generated images of real, identifiable people without consent.
• Using AI to launder copyright-infringing material by paraphrasing it superficially.

Deception

• Using AI to impersonate another editor, administrator, or functionary.
• Using AI to create fake consensus, fake votes, or fake support in discussions.
• Using AI-generated images or audio to misrepresent events, people, or places in article content.
• Concealing AI involvement where disclosure is required by this policy.

User responsibility and accountability

Every edit published on Bharatpedia is the responsibility of the editor who saves it. This is true whether the edit was typed by hand, produced with AI assistance, or copied from another source.

The following claims are not valid defences:

• "The AI wrote it, not me."
• "I did not check the output because the AI seemed confident."
• "The tool told me this was correct."
• "I assumed the model had access to reliable sources."
• "I did not know the image was AI-generated."
• "The AI gave me the citation, so I assumed it existed."

Editors who publish AI-assisted content remain fully accountable for:

• accuracy of facts;
• verifiability of sources;
• compliance with copyright;
• compliance with biographies of living persons policy;
• compliance with neutral point of view;
• civility and conduct in discussions;
• licensing of media files.

Sanctions for violations of this policy are applied to the human editor, not to the AI tool or its developer.

AI in discussions

AI may assist editors in preparing for discussions, but it must not replace human judgement.

• Editors may use AI to organise their own thoughts, draft a comment, or check their reasoning before posting to avoid grammatical and punctuation errors.
• Editors must not post completely AI-generated comments wholesale in deletion discussions, requests for comment, arbitration, or noticeboards. They should express their own ideas and opinion.
• Editors who are unable to write in English, or who struggle to express their thoughts in English, may use AI to phrase their comment. In such cases, they must provide attribution in the edit summary identifying the AI tool used.
• AI-assisted comments that misrepresent another editor's position, or that fabricate quotations from a discussion, are prohibited.

Disclosure requirements

Disclosure is required when AI has materially contributed to an edit. "Materially" means the AI shaped the substance of what was published, not merely the surface.

Disclosure is required for:

• any AI-generated image, audio, or video file uploaded to Bharatpedia;

Disclosure is not required for:

• spelling and grammar fixes;
• punctuation and capitalisation corrections;
• formatting, template, and category fixes;
• reference syntax repairs;
• minor wording improvements that do not change meaning.

AI-generated media

AI-generated or AI-manipulated images, audio, and video are subject to strict rules.

• AI-generated images may be used only where they illustrate a concept, not a real person, place, or event.
• AI-generated images of real, identifiable people are prohibited in article space.
• AI-generated images of historical events are prohibited in article space.
• AI-generated images must be clearly tagged on their file description page with an appropriate template.
• AI-upscaled or AI-restored versions of historical photographs must be labelled as such, and the unmodified original should be preserved where possible.
• AI-generated audio and video are subject to the same rules and must be clearly labelled.

AI-generated media must comply with non-free content policy and image use policy where applicable. A free licence claim on an AI-generated file does not excuse misrepresentation.

Copyright and licensing of AI output

The copyright status of AI-generated output is unsettled in many jurisdictions, including India. Editors should therefore exercise caution.

• Do not assume that AI output is free of third-party copyright. Models are trained on large corpora and may reproduce protected expression.
• Do not upload AI output while claiming it as your own original work if you did not create the underlying expression.
• Do not use AI to paraphrase copyrighted text and then present the result as original writing.
• Do not use AI to generate images that closely imitate a living artist's distinctive style for use in article space.

Where AI output is used, the editor must ensure that the final published version does not infringe copyright and complies with Bharatpedia's licensing requirements.

AI in article creation

New articles are the highest-risk area for AI misuse. The following rules apply.

• Do not create a new article by pasting AI output and publishing it without verifying the content and sources manually.
• Do not use AI to expand a stub into an article without verifying each added claim.

Editors who repeatedly create AI-generated articles that fail verification may face editing restrictions.

Detection and review

Bharatpedia does not rely on AI-detection tools as the sole basis for sanctions. Detection tools are unreliable and produce both false positives and false negatives.

However, editors, administrators, and functionaries may consider the following indicators:

• claims that cannot be traced to any real source;
• images with tell-tale AI artifacts;

Where AI misuse is suspected, the usual dispute resolution processes apply. Administrators should avoid accusatory language and should focus on the substance of the concern accuracy, sourcing, licensing rather than on speculation about tools.

Sanctions

Violations of this policy are handled under existing Bharatpedia policies on disruptive editing, vandalism, copyright violations, biographies of living persons, and personal attacks.

Possible consequences include:

• a warning or advisory notice on the editor's talk page;
• a request to disclose AI use;
• removal or revision-deletion of AI-generated content;
• deletion of AI-generated media files;
• a topic ban from article creation, translation, or discussion venues;
• a temporary or indefinite block, in serious or repeated cases;
• referral to arbitration where the conduct is part of a broader pattern.

Sanctions apply to the editor, not to the AI tool. Editors cannot transfer responsibility for their edits to a software provider.

Bots, gadgets, and automated tools

Bots, scripts, and gadgets that incorporate AI models are subject to the existing bot policy and the following additional requirements:

• the operator must describe the model, its purpose, and its limitations;
• the bot must not publish AI-generated prose to article space without human review;
• the bot must not generate or alter citations;
• the bot must log its AI-assisted actions in a reviewable manner;
• the operator remains accountable for every action taken by the bot.

AI and the Foundation

Nothing in this policy limits the ability of the Bharatpedia Foundation, its Board, or system administrators to take action on AI-related matters under office actions or applicable law. Where this policy conflicts with a mandatory provision of Indian law, the law prevails.

This page is a Bharatpedia community policy and is not legal advice.

Review and amendment

This policy should be reviewed periodically, at least once every twelve months, given the pace of change in AI technology and law. Amendments should be proposed on the talk page and adopted through the normal consensus process.

Editors are encouraged to propose additional permitted uses, new categories of prohibited use, and updates to disclosure requirements as the technology develops.

See also

• Bharatpedia:Copyrights
• Bharatpedia:Verifiability
• Bharatpedia:No original research
• Bharatpedia:Biographies of living persons
• Bharatpedia:Non-free content
• Bharatpedia:Bots
• Bharatpedia:Disruptive editing
• Bharatpedia:Plagiarism
• Bharatpedia:Identifying reliable sources
• Bharatpedia:Deletion policy
• Bharatpedia:Artificial intelligence in education
• Bharatpedia:Machine translation