Nexus Mods AI Rules Explained: Disclosure Is Now Mandatory — But There’s No Ban

Key Takeaways

• Nexus Mods announced new generative-AI rules on 3 November 2026 in a site news post, “AI-Generated Content on Nexus Mods”, by site owner Victor “Foledinho” Folmann. It is not a ban on AI content.
• Every mod author must now declare whether generative AI was used, choosing one of three labels: AI-Generated Content, AI Media, or AI-Assisted.
• AI-Assisted mods must also carry a public explanation of how and where AI was used, including the tools used and which assets they produced.
• Players can hide AI-tagged mods using content-preference controls on game pages and the homepage, with search filtering to follow.
• Existing mods have a 60-day grace period, running from 3 November 2026 to 2 January 2027, to comply. After that, missing disclosures can lead to warnings, account restrictions or removal.
• Nexus says it will not use uploaded mods to train AI models, and will not sell or license them to anyone who does.
• Search interest in “nexus mods artificial intelligence rules” is listed as Breakout in the US and the UK over the past seven days — a measure of rising attention, not of how many people support or oppose the rules.

What Nexus Mods announced

On 3 November 2026, Nexus Mods published a site news post titled “AI-Generated Content on Nexus Mods”, written by site owner Victor “Foledinho” Folmann. The post sets out new rules for generative AI on the platform — and the first thing to understand about it is what it does not do. It does not ban AI-generated mods.

That distinction is the whole story, in a sense. Nexus Mods is one of the largest modding platforms in the world, hosting mods for hundreds of games, and generative AI has been arriving on the platform for some time — in code, voices, textures, thumbnails and descriptions. The community has been arguing about it for just as long. The new rules are Nexus’s answer to that argument: not a ban, but a system of mandatory disclosure, labelling and filtering, backed by a promise not to use mods for AI training.

Folmann set out the principle behind the rules in a single line: “We are pro-human creativity. That’s the basic premise behind all decisions below”. And he was equally direct about the outcome: “AI-generated mods are welcome on our platform provided that they meet certain criteria regarding permissions, transparency, presentation, and conduct.” The rules that follow are those criteria, spelled out.

The announcement also comes with an admission that the policy will not satisfy everyone. Folmann said he does not expect everyone to agree, and that the policies will keep evolving. That is worth keeping in mind throughout this article: these rules are a starting position, not a final settlement, and the reaction so far suggests the debate will continue.

The three labels every mod author must now choose

The centrepiece of the new rules is mandatory disclosure for new uploads. Every mod author must now declare whether generative AI was used in their mod, and if so, how. There are three labels, and the choice between them depends on how central AI was to the work.

The first label is AI-Generated Content. This applies where AI produced most of the code, voices, textures or other core assets of the mod — in other words, where the substance of the mod itself is largely AI-made. The second label is AI Media. This covers cases where the mod itself may be human-made, but promotional material was created with AI: the art, thumbnails or page text used to present it. The third label is AI-Assisted. This is for human-led work where AI played a supporting role — the examples given are texture upscaling and audio cleanup — helping a human author rather than doing the core creative work.

The distinction between the three matters because they signal very different things to a player browsing the site. A mod labelled AI-Generated Content tells you that the thing you are about to download was mostly produced by a machine. A mod labelled AI-Assisted tells you that a human made it, with AI tools helping on specific tasks. And AI Media tells you something else again: that the mod may be entirely human-made, but the picture or description selling it to you was not. Lumping those three together would mislead players in different directions, and the three-label system is Nexus’s attempt to avoid that.

There is also an extra requirement for the AI-Assisted label. Mods carrying it must include a public explanation of how and where AI was used, including the tools used and which assets they produced. That goes beyond a tick-box: an author cannot simply select AI-Assisted and move on. They have to say, in public, what the AI did and what it touched. For players who want to make an informed choice, that explanation is arguably more useful than the label alone, because it shows the boundary between the human work and the machine’s contribution.

The labels will also be more visible than before. AI tags will be shown more prominently on mod pages, placed right after the adult-content tags — a position that signals Nexus considers them important information, not a footnote. That placement also has a practical effect: players will see the AI status of a mod early, before they read the description or look at the screenshots, rather than having to hunt for it.

This is part of a wider shift towards labelling AI-generated material so that audiences can tell what they are looking at. A similar instinct is visible elsewhere in the AI world: our report on the Grok API Watermark Starts 7 November: What Builders Must Do covers a labelling system of a different kind — watermarking AI-generated content at the point of creation — but the underlying idea is the same. When audiences cannot tell human-made from machine-made work, trust breaks down; labelling is the attempt to restore it.

What players will see — and what they can hide

The new rules are not only about what authors must declare. They also change what players can control. Nexus is adding content-preference controls on game pages and the homepage that let players hide AI-tagged mods, and search filtering is to follow. In practice, that means a player who does not want to see AI-generated mods will be able to filter them out of the listings they browse, rather than having to check each mod individually.

That is a significant concession to the part of the community that wanted a ban, without being a ban. A ban would remove AI mods from the platform for everyone; filtering removes them only for the players who choose to hide them. The difference matters for authors as well as players: an AI-assisted modder’s work remains available to the players who want it, while players who prefer human-made mods can browse without encountering it.

It is also worth noting what has not yet arrived. Search filtering is described as following the page-level controls, not launching with them. So the filtering experience will, at least at first, be partial: available on game pages and the homepage, with search to catch up. Players who rely mainly on search to find mods will need to wait for that piece, and authors should expect the visibility of AI-tagged mods in search to change when it does.

What happens to existing mods: the 60-day grace period

The new disclosure requirement applies to new uploads immediately, but the millions of mods already on the platform are not expected to comply overnight. Existing mods have a 60-day grace period, running from 3 November 2026 to 2 January 2027, to add their disclosures.

After that period ends, authors will be asked to fix missing disclosures or permissions problems. Non-compliance can lead to warnings, account restrictions or removal — a graduated sequence, not an immediate purge. The emphasis, at least as described, is on getting authors to fix the problem rather than on punishing them for it, with removal as the end of the scale rather than the start.

Enforcement will be manual, and Nexus is candid about its limits. Moderation will rely heavily on player reports, and Nexus admits that spotting AI content “is by no means an exact science”. That is an honest statement, and an important one, because it sets expectations for how the rules will work in practice. AI-generated textures, code and voices are not always distinguishable from human-made ones by inspection, and tools that claim to detect AI content are unreliable. A system that depends on player reports will therefore catch some undisclosed AI use and miss other cases, and it will sometimes generate disputes about whether a particular mod was AI-assisted or not.

That uncertainty is also why the public explanation requirement for AI-Assisted mods matters. When a mod’s AI use is declared openly, with tools and assets named, there is less for moderators and players to argue about. The rules work best where authors comply voluntarily, and the grace period is, in effect, Nexus asking them to do exactly that — to go back through their existing uploads and label them before enforcement begins.

The no-training promise

One of the most consequential parts of the announcement has nothing to do with labels or filters. Nexus says it will not use uploaded mods to train AI models, and will not sell or license them to anyone who does. For mod authors, that is a direct answer to one of the most persistent fears about hosting creative work on a large platform: that the work will be harvested to train the very systems that then compete with its creators.

That promise sits alongside new permissions options, which are still in progress. The new options will cover AI usage, AI training, forks and derivative mods, and they will default to “Ask me” — meaning that, unless an author chooses otherwise, anyone wanting to use their work in those ways will have to ask first. Existing permissions will not be reset, so authors who have already set their preferences will not find them silently changed. And content that uses another author’s work without permission will not be hosted.

The permissions question connects to a much larger debate about AI training data and creators’ rights, which is playing out well beyond gaming. Our coverage of AI Training Opt-Out: One Rule for Four AI Labs looks at how major AI labs handle opt-outs from training, and the EU AI Copyright Consultation Closes 3 November: What Creators and AI Firms Must Know covers the consultation in Brussels on the same underlying question: who gets to decide whether creative work can be used to train AI. Nexus’s answer, for its own platform, is that the author decides — and that the default is to ask.

Thumbnails and misleading images

One of the most argued-about corners of the new rules concerns thumbnails. Nexus has not imposed a blanket ban on AI thumbnails. An author can use an AI-generated image to present their mod. What is against the rules is misleading imagery — including AI images passed off as gameplay screenshots.

The distinction is about honesty in presentation, not about the tool used to make the image. A thumbnail that is plainly promotional art is one thing; a screenshot that purports to show what the game looks like with the mod installed, but was in fact generated by an AI, is another. The second case misleads players about what they are downloading, and it is that deception the rule targets.

This is also the area where community frustration has been loudest, as the reaction below describes. Some players feel that AI thumbnails have already degraded the browsing experience, and that a rule against only the misleading ones does not go far enough. Nexus’s position is that the problem is misrepresentation, and that an honest AI thumbnail is acceptable. That is a line that will be tested in practice, case by case, by moderators and player reports.

Donation Points: unchanged, and a revealing number

Nexus has not changed its Donation Points system for now. Donation Points are the mechanism through which mod authors can earn money from their work on the platform, and the question of whether AI-generated mods should earn them at the same rate has been part of the community debate.

The announcement included one figure that puts that debate in context: mods tagged AI-Generated Content received 2.13% of all Donation Points so far this year. Nexus called that “a comfortably low percentage”. The implication is that AI-generated mods are not, at present, taking a large share of the money — which is why Nexus feels able to leave the system unchanged for now. It is also a figure worth watching: if the share rises sharply, the question of whether AI mods should earn Donation Points on the same terms will almost certainly come back.

Why now: the security context

The AI rules were announced alongside a security context that helps explain their timing. Over the last year, Nexus quarantined 20,000 uploads for extra checks — 12 times the number in the year before. And between November 2025 and October 2026, 3,164 uploads were manually reviewed and cleared.

Those numbers are not directly about AI disclosure, but they are about the same underlying pressure: the volume and the nature of uploads to the platform are changing, and the systems that check them are under strain. AI tools make it easier to produce mods quickly, and easier to produce large numbers of them, including low-effort or malicious uploads. A quarantine system that handled a certain volume a year ago is now handling 12 times as much, and manual review — 3,164 uploads cleared in a year — is labour-intensive by definition.

The disclosure rules can be read as part of the response to that pressure. If authors must declare their AI use, and if players can filter and report, the moderation burden is shared more widely, and the platform has more information about what it is hosting. It is not a complete solution to the volume problem, and Nexus does not present it as one, but it is a step towards making the platform’s contents more legible — to players, to moderators, and to Nexus itself.

A divided community

The reaction to the announcement has been divided, and the split is instructive. As paraphrased from VGC’s coverage, some members of the community want an outright ban on AI-generated content and are frustrated by AI thumbnails; others say that generative AI and AI-assisted coding made modding more accessible, and feel targeted by the new rules.

Both positions have a logic to them. For the first group, the concern is about the character of the platform: that a space built on human creativity is being filled with machine-generated work, and that labelling and filtering are not enough to protect it. The frustration about thumbnails is part of the same concern — that even the browsing experience, the first thing a player sees, is increasingly AI-made.

For the second group, the concern is about access. AI-assisted coding and AI tools have lowered the barrier to making mods, allowing people without years of programming or art experience to build things they could not have built before. From that perspective, the new rules can feel like a penalty for using the tools that made modding possible for them — and the label itself, prominently displayed, can feel like a mark against their work.

Folmann addressed the disagreement directly, in a line that captures the difficulty of Nexus’s position: “An outright ban will not eliminate AI from the modding process but will prevent people from mentioning its use.” The argument is practical rather than philosophical. A ban, in his view, would not remove AI from modding — it would drive it underground, where it could not be labelled, filtered or moderated. Disclosure, by contrast, keeps it visible, and visibility is what makes the rest of the system work.

It is also worth noting that Nexus is not the only gaming organisation grappling with this question. Our report on the UK Games Expo AI Policy: What the New Rules Mean for Exhibitors in 2027 covers another set of AI rules in the gaming world — for exhibitors rather than modders — and the Saber Interactive AI Disclaimer: What It Actually Says (2026) looks at how a game developer discloses its own AI use. The pattern across all three is the same: disclosure and labelling, rather than prohibition, as the chosen tool — and the same debate, in each case, about whether that is enough.

What mod authors should do now

For authors, the practical implications are straightforward, and the grace period gives them time to act. First, go through existing uploads and decide which of the three labels each one needs — AI-Generated Content, AI Media or AI-Assisted — remembering that a mod can carry a label for its promotional material even if the mod itself is human-made. Second, for anything labelled AI-Assisted, write the public explanation: which tools were used, and which assets they produced. Third, check permissions, bearing in mind that the new options default to “Ask me” and that content using another author’s work without permission will not be hosted. Fourth, check thumbnails and screenshots: AI thumbnails are not banned, but anything that could be read as a gameplay screenshot when it is not one should be fixed. And fifth, do it before 2 January 2027, when the grace period ends and enforcement begins.

For players, the practical change is simpler: the content-preference controls on game pages and the homepage will let you hide AI-tagged mods, with search filtering to follow. The choice is yours, and it is reversible — which is, in the end, the point of the system Nexus has built. Rather than deciding for everyone whether AI mods belong on the platform, it is asking each player to decide for themselves, and asking each author to be honest about what they made.

The bottom line

Nexus Mods has chosen disclosure over prohibition. Every mod author must now say whether generative AI was used, and how; players can hide the AI-tagged mods they do not want to see; existing mods have until 2 January 2027 to comply; and the platform has promised not to use mods for AI training or to license them to anyone who does.

Whether that settlement holds will depend on three things. The first is compliance: whether authors label their work honestly during the grace period, or whether the rules have to be enforced case by case after it. The second is moderation: whether a manual system, relying heavily on player reports, can keep up with a platform where AI content is, by Nexus’s own admission, not an exact science to spot. And the third is the community itself: whether the two sides of the argument — those who wanted a ban and those who feel targeted — accept a system that gives neither of them everything they asked for.

Folmann said he does not expect everyone to agree, and that the policies will keep evolving. On the evidence of the reaction so far, both halves of that sentence are true. The rules are a beginning, not an end — and the next chapter will be written by how authors, players and moderators use them.

Sources: VGC, “Nexus Mods rules out full AI ban despite community backlash”; Inven Global, “Nexus Mods makes AI disclosure mandatory for every mod but won’t ban AI content”; TweakTown, “Nexus Mods now requires creators to disclose AI use but isn’t ready to ban AI-generated mods”; Nexus Mods news.