Nexus Mods, the largest PC game modding site, said on November 4, 2026 that authors of newly uploaded mods must declare whether AI was used in their work. Mods tagged AI-assisted must also carry a public explanation of how and where AI was used. The site will not ban AI mods. Existing mods have 60 days to comply, until January 2, after which enforcement starts.
The announcement comes from site owner Victor "Foledinho" Folmann in a longer post on AI-generated content, as reported by The Verge. We could not open Nexus Mods' own page, which returned an access error to our fetch tool, so everything about the November 4 post is as The Verge reports it. The earlier tag system, from late July, is described by Shacknews.
What exactly is changing? #
| Change | Detail (per The Verge) | Status |
|---|---|---|
| Mandatory declaration | Authors of new mods declare "whether AI has been used in their work" | Applies to newly uploaded mods |
| Explanation | Mods tagged AI-assisted need a public explanation of how and where AI was used | Required |
| Tag prominence | AI tags more visible on mod pages | Rolling out |
| Hide AI content | Controls on game pages and the homepage; search "next up" | Launching, search pending |
| Existing mods | 60 days to update | Grace period ends January 2 |
| Training use | Site will not use mods to train AI, or sell or license them to anyone who does | Stated policy |
| Reuse and attribution | "Clearer control" over reuse of mods, stronger attribution for mods built on others | In development |
| Recently Added | Limits on how many mods per author show at once; tighter limits on new accounts | Being tried |
Why not a ban? #
Folmann's reasoning, quoted by The Verge: "AI is currently used throughout the whole process of creation, from script autocompletion to generation of textures, voices, and full assets, and it is difficult to draw a clear boundary between what is acceptable and unacceptable. An outright ban will not eliminate AI from the modding process but will prevent people from mentioning its use."
That is a disclosure-over-prohibition argument, and it is the same one platforms such as Steam and itch.io have used with different results. Our coverage of Steam's AI disclosure and the missing no-AI filter and itch.io deleting terabytes of AI slop shows both the appeal and the limits of that route. Nexus Mods is choosing the filter option that Steam lacked, by letting players hide AI-tagged content.
A game controller with a green spark, representing the modding community whose AI rules Nexus Mods is rewriting
What came before: the July tag split #
On July 31, 2026, Nexus Mods split its single "AI-Generated Content" tag into three, according to Shacknews:
- AI-Generated Content: AI-generated code, user interface elements, voices, dialogue, translations, music and in-game assets.
- AI Media: AI promotional art, thumbnails, videos, mod page descriptions or media outside the mod content in game.
- AI Assisted: developer-led code, content or assets with limited AI involvement, such as AI texture upscaling, audio upscaling or cleaning, and AI concept art not in the final mod.
Shacknews adds that failure to tag correctly would bring moderation action, and that deliberate abuse of the tagging system would be punished immediately if proven. The November 4 step turns the tags from something authors could choose into something they must declare, and adds the explanation requirement for AI-assisted work.
Why does the rule matter beyond games? #
Modding is a test case for a hard problem: how to label AI use when it is partial. A mod might use AI-upscaled textures, human-written code with autocomplete help, and a voice line cloned from an actor's recordings. A yes/no label cannot carry that.
Nexus Mods' answer is a three-tag taxonomy plus a free-text explanation. Whether authors write honest explanations is unverified. No detector is mentioned in The Verge's report, so enforcement rests on community reports and moderator judgment. That is the weak point of every disclosure regime, and the reason the January 2 enforcement date is worth watching.
Folmann says the feedback driving the change was consistent: users wanted to know when AI had been used, and said so was not clearly disclosed. The same post deals with upload flooding. The Verge notes the Recently Added changes address upload flooding that is "less of an issue related to AI" but growing as mod production speeds up. For a parallel on search pollution, see our post on AI-generated fan wikis flooding game search results.
How does it compare with other platforms? #
Nexus Mods combines three things in one policy: a mandatory declaration for new uploads, a free-text explanation for AI-assisted work, and a player-side control to hide AI content. Our Steam and itch.io posts above describe platforms where disclosure and filtering developed separately, and where the lack of a player filter was a recurring complaint. Read those posts for the exact state of each store, because platform rules change often.
Which tag would a real mod get? Worked examples #
The tag definitions come from Shacknews' account of the July guidelines. The examples below are our reading of them, not Nexus Mods rulings, so treat them as a starting point and check the site's policy text.
| Mod | Likely tag | Explanation to write |
|---|---|---|
| New quest with dialogue lines written by a language model | AI-Generated Content | Which lines, which tool, and whether a human edited them |
| New voice lines made with a cloned voice | AI-Generated Content | The source of the voice and the permission status |
| Texture pack upscaled with an AI upscaler from original game assets | AI Assisted | The upscaler, the settings, and the fact that the base art is the game's |
| Hand-written script mod, with editor autocomplete suggestions | Possibly none, or AI Assisted | Where autocomplete was used, if you count it |
| Hand-made mod with an AI-generated thumbnail and description | AI Media | That only the page media was generated |
| Mod where a model wrote the whole script | AI-Generated Content | The tool and the testing you did yourself |
The edge cases are the interesting ones. Folmann's own words show why: AI is used through the whole creation process, from script autocompletion to full assets. Where one tool is used for a trivial edit, the line is a judgment call. The explanation field exists to let an author state that judgment openly.
What are people asking about the rule? #
Does a disclosure label mean a mod is bad? No. Labels sort content, they do not rate it. A texture mod that uses AI upscaling may be excellent. The point is that players can choose.
Can the site verify the labels? Not by the evidence in the coverage we read. Enforcement starts after January 2, and the penalties for mis-tagging come from the earlier July guidelines, which Shacknews says include moderation action and immediate punishment for proven abuse of the tags.
What about mods that were uploaded years ago? The 60-day window applies to existing mods. Authors who are inactive may not update, which leaves a gap that the site will have to handle. The coverage does not say how.
Will this reduce the flood of low-effort uploads? The Recently Added changes target volume directly, with per-author limits and tighter rules for new accounts. Those rules apply whether or not a mod uses AI, which suggests the site sees volume and AI as related but separate problems.
Is this a response to law or to users? Folmann frames it as user feedback. Nothing in The Verge's report links it to a regulation, and AI labelling laws elsewhere are a separate topic.
What does this mean for the wider AI-content debate? #
Three lessons show up. First, platforms are converging on disclosure plus filters rather than bans, because a ban pushes AI use out of sight. Second, partial AI use needs a taxonomy, not a checkbox, and Nexus Mods now has three tags plus an explanation. Third, the hardest part is not the policy but the enforcement, which relies on moderators and community reports until better tools exist.
For builders of AI tools, the practical takeaway is that provenance metadata will matter. If your tool exports assets that end up in mods, games or stores, adding an optional machine-readable note about generation helps authors comply with rules like this one. For readers who follow the creative side of AI, our coverage of AI in film and actor consent fights shows the same question of consent and credit in other media.
What should mod authors do now? #
- Audit your mods. List where AI touched code, art, voice, text and promotional images.
- Choose the right tag. Use AI-Generated Content for in-game assets and code made by AI, AI Media for promotional material, and AI Assisted for limited help.
- Write the explanation before the deadline. For AI-assisted mods, say what the tool did and what you did yourself.
- Check voice and likeness rights. If a voice is cloned, get permission. Disclosing AI use does not by itself make unauthorised cloning acceptable.
- Update by January 2. After that date the site starts enforcing.
What should players and server admins do? #
- Use the hide controls once they reach your game page, and check search filtering when it arrives.
- Treat the tag as a claim, not proof. A missing tag does not guarantee human-made content.
- Report mismatched tags. The system depends on it.
What is still unknown? #
- How Nexus Mods will detect false declarations. The Verge's report mentions none.
- When search filtering ships. It is described only as "next up".
- Whether authors of already-published mods who do not update will be unlisted or only tagged.
- How the new reuse and attribution controls will work.
We will update this post if Nexus Mods publishes details.
Related reading #
Details reflect The Verge and Shacknews reporting as of November 4, 2026.