Updated 4 November 2026 — the Nexus Mods AI policy was announced on 3 November 2026, and enforcement begins 2 January 2027.
The biggest modding site in PC gaming has drawn a line on artificial intelligence. Under the new Nexus Mods AI policy, every mod uploaded to the platform must declare whether generative AI was used to create, edit, or process anything in it — no more silent AI-generated mods hiding among hand-made ones.
The change lands with a 60-day grace period, permission defaults that put AI training behind an ‘Ask Me’ wall, and feed fixes aimed at mass-produced mods. Here is what the Nexus Mods AI policy really says, why the site refused an outright ban, and what it means if you make or install mods.
- What The Nexus Mods AI Policy Actually Changes
- Why Nexus Mods Chose Disclosure Over An AI Ban
- The Three AI Labels Under The New Nexus Mods AI Policy
- Permissions, Training Data And The ‘Ask Me’ Default
- Flood Controls: Cleaning Up The Recently Added Feed
- How The Modding Community Is Reacting
- What The Nexus Mods AI Policy Means For PC Gaming
- References
What The Nexus Mods AI Policy Actually Changes
What the announcement actually says
Nexus Mods — the largest home for PC game modifications, run by UK studio Black Tree Gaming — announced on 3 November 2026 that AI disclosure is mandatory across its platform. Under the Nexus Mods AI policy, any author using generative AI to create, edit, or process a file, asset, or piece of page documentation must declare it accurately at upload time.
The deadline, in dates
A 60-day grace period runs from 3 November 2026 to 2 January 2027, and the Nexus Mods AI policy grace period covers only the newest requirements: detailed AI-assisted statements and the new permission tags. Basic disclosure was already site policy. After 2 January, every hosted mod must comply, and non-compliant files go to the Trust & Safety team for review. Repeat failures can end in account restrictions or removal.
Why this landing is worth attention
Nexus Mods hosts hundreds of thousands of files across thousands of games and sits at the centre of PC modding culture. When the platform with the deepest reach in the space formalises its rules, smaller sites tend to follow. That is what makes the Nexus Mods AI policy a bellwether for creator policy well beyond games — and why the reaction has been so loud.
The table below summarises what the Nexus Mods AI policy changes for the average author.
| Area | Before November 2026 | After 2 Jan 2027 |
|---|---|---|
| AI tagging | Optional and easy to miss | Mandatory at upload, shown prominently |
| AI-assisted details | No itemised statement required | Public list of tools, outputs, and evidence |
| AI-training permission | Unspecified by default | Every file defaults to ‘Ask Me’ |
| Forks and repacks | Loose community norms | Explicit authorisation required |
| Recently Added feed | Unlimited posts per author | Per-author display caps, stricter new accounts |
What the change asks of a mod author in practice
For authors, the new workflow is concrete rather than abstract. When uploading, a creator selects the category that fits the work: fully machine-made material, machine-made media sitting beside human-built gameplay, or an ordinary mod built with machine help somewhere in the process.
The third path carries the most paperwork. An AI-assisted declaration should name the tools used, describe which outputs they produced, and point to supporting evidence such as development history or source commits, so reviewers can check the claim instead of taking it on faith.
Authors also need to visit the permission panel once and settle two questions: whether their files may train future models, and whether other users may fork them. Both default to a prompt rather than a silent yes. Existing uploads fall under the same audit, with the grace window running into early January 2027 before Trust and Safety starts issuing strikes.
Why Nexus Mods Chose Disclosure Over An AI Ban
The boundary problem
Writing off AI wholesale sounds easy until you define what counts. Nexus Mods owner Victor ‘Foledinho’ Folmann put it plainly: 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.’ A script that autocompletes a function is mundane; an AI-generated voice pack for a beloved RPG is not — yet both are AI under the Nexus Mods AI policy.
Sunlight as the regulation of choice
The Nexus Mods AI policy bets on disclosure rather than prohibition. ‘An outright ban will not eliminate AI from the modding process but will prevent people from mentioning its use,’ Folmann wrote. ‘We’d prefer the opposite; it should be out in the open, clearly labelled, and regulated according to the rules that we can implement on our platform.’ Visible labels, the site argues, beat underground usage.
What a ban would have cost
A prohibition would have turned away authors who use AI tools casually — the same hobbyists who keep decade-old RPGs alive. It would also have been unenforceable: moderators cannot open every mesh or texture and check its provenance. Labelling is the only mechanism a platform can actually police, which is why it anchors the entire Nexus Mods AI policy.
The Three AI Labels Under The New Nexus Mods AI Policy
Class one: AI-generated content
Authors pick this label when AI produced the majority of the mod’s core work — code, UI, voices, dialogue, translations, music, or other assets. AI-generated textures, synthetic voice lines, and generated scripts all land here. It is the headline class of the Nexus Mods AI policy and carries the simplest obligation: say so, truthfully, at upload.
Class two: AI media
This catches the packaging rather than the mod: AI promotional art, thumbnails, videos, and mod page descriptions — including AI upscaled or enhanced images. A hand-built quest mod fronted by an AI-generated thumbnail is an AI media case, so even cosmetic choices fall under the Nexus Mods AI policy.
Class three: AI-assisted, the strictest bucket
For human-led work with limited AI involvement — texture upscaling, audio cleaning, concept art that never ships — authors select AI-assisted, and the bar rises. The mod page must list the specific tools, software, or models used, and state exactly what AI outputs survive into the final mod, from autocompletion to dialogue polish. Authors must also be ready to show development history, version-control commits, and genuine understanding of their code in any moderation dispute. Claim the label without evidence and moderators may reclassify the mod under the Nexus Mods AI policy.
Here is how the three classes compare under the Nexus Mods AI policy:
| Label | What it covers | Core obligation |
|---|---|---|
| AI-Generated Content | AI made the majority of code, UI, voices, dialogue, translations, music, or assets | Disclose accurately at upload |
| AI Media | Promo art, thumbnails, videos, page descriptions, AI-upscaled images | Disclose on the mod page |
| AI-Assisted | Human-led work with limited AI support | List tools used, identify AI outputs, keep evidence |
Worked examples: which label fits which mod
Imagine a texture pack in which every surface came from a generative model with no hand-authored source. That falls under the first class: the content itself is machine-made, and the label sits front and center.
Now picture a hand-built overhaul whose author commissioned a machine-made cover image for the listing while every script, mesh and quest remains human work. That is media, not content: gameplay stays untouched, and the disclosure covers only the promotional material.
Finally, consider a companion mod whose author wrote the logic by hand but asked a coding assistant to generate a handful of utility functions, then committed them alongside their own changes with notes in the changelog. That is assistance. The statement should name the assistant, say which functions it produced, and let the commit history back the story up.
Permissions, Training Data And The 'Ask Me' Default
The ‘Ask Me’ default explained
Alongside labelling, the Nexus Mods AI policy introduces explicit permission settings for two things authors have complained about for years: AI training and forking. Every hosted file — including legacy mods published long before these options existed — now defaults to ‘Ask Me’. For anyone uploading, that means choosing AI-training and fork preferences under the Nexus Mods AI policy before a new file goes live.
No training on your mods without written permission
‘Ask Me’ is not a polite suggestion. It strictly prohibits anyone using a mod’s meshes, textures, scripts, audio, text, or source code for training, fine-tuning, or dataset compilation, or for feeding files into automated AI editing workflows, without direct written permission. Nexus Mods has repeated that it will not use mods to train AI itself and will not sell or license them to anyone doing so — a platform-level opt-out comparable to the controls we covered in Google’s new AI opt-outs for schools.
Attribution gets real teeth
Forks, parallel versions, patches, and modified assets must credit every original creator and link back to the source mod — in the main description, through the full chain of derivative works, with working links even when the source lives on another site. One nuance survives even this: copyleft licences such as GPL still outrank platform defaults. For authors who watched their work repackaged for years, this credit chain is the quiet heart of the Nexus Mods AI policy.
Flood Controls: Cleaning Up The Recently Added Feed
The Recently Added problem
Flooded discovery feeds are the collateral damage of fast production. Folmann concedes the spam is ‘less of an issue related to AI’ but admits it is ‘becoming much more prevalent as mod production is speeding up.’ One author dumping fifty files in a sitting pushes everyone else off the front page — which is why the Nexus Mods AI policy package ships with feed changes, not just tags.
The new feed rules
Under the Nexus Mods AI policy overhaul, Recently Added will cap how many mods from a single author appear at a time, and new accounts face tighter upload restrictions from day one. The goal is simple: discovery should reward steady, human-paced releases instead of bulk uploads timed for the front page. The policy treats feed hygiene as part of AI governance, because volume is how unlabelled floods hide.
Player-side hiding controls
Players were never asked whether they wanted AI content served to them; now they can decline it. Controls to hide AI-generated content rolled out on game pages and the homepage alongside the Nexus Mods AI policy launch, and the site says search filtering is ‘next up.’ It is a quiet admission that discovery defaults, not just tags, shape what players see — and that defaulting to showing everything was itself a choice.
Why feed health matters more than it sounds
A mod site lives or dies on discovery. Players browse Recently Added to find something to play this weekend; authors post updates hoping exactly those browsers will notice. When the stream clogs, both sides lose: players skim past good work, and quiet hobbyists stop getting downloads at all.
That is why the feed changes accompany the labels instead of arriving later. Clearer metadata gives the browsing layer something to sort on, and it gives readers a reason to trust what surfaces at the top. Framed this way, the labelling rules are less a culture war than plumbing repair: the site is protecting the browse experience that keeps its community logging in.
How The Modding Community Is Reacting
Why many modders approve
For artists and coders who watched AI-generated ‘mods’ clog queues over the past two years, mandatory labels feel like overdue housekeeping. The Nexus Mods AI policy’s no-training default is the bigger win: it converts the community’s loudest demand — keep our work out of training sets — into an enforceable platform setting rather than an honour system.
Where modders worry
The evidence requirements sting. Proving human-led development with commit history sounds reasonable until you are a hobbyist who works in loose files. Some fear the Nexus Mods AI policy dispute process will burden accused authors rather than flaggers, and that under-resourced moderators will get the worst of both worlds. Others note that AI-assisted covers so much — autocomplete included — that the label may blur rather than inform.
The enforcement test
Every policy is judged after the grace period, and the Nexus Mods AI policy names real consequences: staff notification, moderation action, account restrictions, removal. It leans on community flagging to find violations at scale. Whether Trust & Safety can process thousands of borderline files is the open question the whole industry will watch through January.
Where the critics come in
Not everyone reads the policy as protection. Some authors argue that a disclosure line changes nothing about material that models were already trained on, scraped portfolios and fan art included, and that labelling a finished product cannot undo the process that produced the tool.
Others worry about overhead: hobbyists who publish on weekends now face category choices, evidence trails and a permissions audit, while the platforms that flooded the market face no comparable obligations. A smaller number have said they will simply stop publishing and host files privately rather than sign up to what they see as normalization.
The counter-argument from the site’s defenders is blunt: without rules, the alternative was an arms race of unmarked machine output, and the community would have lost that fight quietly instead of loudly.
What The Nexus Mods AI Policy Means For PC Gaming
The platform scorecard
Compared across storefronts, the Nexus Mods AI policy is the most detailed AI-disclosure regime any major PC platform has adopted. Steam added mandatory disclosure of generative AI in January 2024, but its rules stop at a submission checkbox plus a store-page notice. itch.io added optional disclosure badges in mid-2025. Nexus Mods goes further: three label classes, itemised explanations, default training opt-outs, fork authorisation, and attribution chains.
| Platform | Rules live | What they demand |
|---|---|---|
| Steam | January 2024 | Disclose generative AI at submission; notice on store page |
| itch.io | Mid-2025 | On/off disclosure badge for generative AI |
| Nexus Mods | Enforced from 2 Jan 2027 | Labels, explanations, training opt-out, fork rules, attribution |
How far each platform goes
The Nexus Mods AI policy spans five distinct rule categories — roughly triple Steam’s disclosure regime — which is why many expect copycat policies across mod hosting.
Signals for creators everywhere
Two lessons travel well beyond modding. First, platforms are choosing radical transparency over bans because transparency is what they can actually enforce. Second, default permissions beat policy prose: a platform-wide ‘Ask Me’ does more for training consent than any statement of values. If you build tools on top of user-created libraries, read those Nexus Mods AI policy defaults carefully — this framework may become the template your own platform copies.
What to watch next
Three things matter between now and enforcement: how fast authors update existing files, whether search-level hiding controls arrive as promised, and whether rival mod sites adopt a Nexus Mods AI policy of their own. For deeper context on the tools at the centre of this debate, see our AI models and tools hub and our overview of artificial intelligence and machine learning. The policy will be judged by its first year — but as a signal of where creator rules are heading, it already lands.
Open questions into 2027
Several things remain genuinely unsettled. Enforcement is the first: reading declarations from hundreds of thousands of existing uploads, and policing new ones, is a large job for a hosting platform, and early strike decisions will set the real tone of the policy.
Portability is the second. Steam already requires disclosure of machine-generated content at submission time, and other storefronts may follow the modding sites’ lead, so an author who builds clean provenance habits now is ahead of the curve wherever they publish next.
The third is cultural: whether disclosure becomes a boring checkbox, the way content ratings did, or stays a live argument. The next wave of policy debates on this site will tell us which.
References
More AI coverage: explore Progressive Robot's AI Models, Tools & Releases hub — hands-on reviews, setup guides and benchmarks in one place.
