ArtStation Turns On Strict AI-Art Filters by Default

What changed with the ArtStation AI filter?

ArtStation expanded its generative-AI controls on November 9, 2026, with creator disclosures, viewer filters, and a reporting system for suspected unlabeled AI artwork. The platform now asks creators to identify a published project as primarily generative AI, mixed generative and non-generative, or 100% non-generative. ArtStation’s announcement describes the update as a way to provide more information about how artwork was made and more control over what viewers see.

ArtStation does not ban AI-made artwork outright. The policy change instead separates disclosure from discovery, meaning creators can still publish work while viewers can filter categories they do not want included in their browsing results. This matters because an illustration portfolio can now communicate more clearly whether a project relied on a generative tool, a mixed workflow, or no generative AI.

ArtStation’s new approach gives users more control, but it does not turn a feed setting into a final judgment about authorship. Creator labels and community reports provide signals for the platform’s filters. The practical response is to use the controls as a way to shape browsing results while recognizing that the underlying classification system has limits.

How do creators label generative-AI artwork on ArtStation?

ArtStation creators now choose among 3 project declarations when publishing work: primarily generative AI, mixed generative and non-generative, or 100% non-generative. The declaration places the responsibility for a project’s initial classification on the person publishing it. That structure gives viewers a clearer starting point than an unlabeled portfolio page, especially when a project includes both hand-created work and generated elements.

Mixed-workflow labeling is important because AI use is not limited to a single type of creative process. A creator may use generative tools for ideation, reference, texture experiments, or image components while completing other parts of a project through non-generative methods. The new declaration categories acknowledge that distinction rather than treating every project as either fully AI-made or fully non-AI.

Creators should select the declaration that best reflects the project’s primary process and remain consistent across their portfolio. Clear disclosure can also help artists who use legitimate digital editing tools explain their process without being placed in the same category as work primarily generated by an AI system. Artists comparing image-generation workflows can also consider how AI image generators differ by use case before deciding how to describe their own process.

What does ArtStation’s default Strict filtering do?

ArtStation’s default configuration hides projects declared AI-generated, projects tagged CreatedwithAI, and projects associated with generative-AI tools. The same default configuration sets suspected-AI content to Strict filtering. Strict is the most aggressive of ArtStation’s 3 suspected-content settings: Strict, Balanced, and Loose.

ArtStation says Strict filtering acts when only a few signals suggest that a project may be AI-generated. Balanced filtering requires a moderate level of reports, while Loose filtering acts only when user signals are strongest. The difference matters because Strict prioritizes removing potentially AI-generated work from a viewer’s results, even when available signals are limited.

Strict filtering offers the strongest default separation for viewers who want to focus on non-generative portfolios. At the same time, the setting can affect work that has not been conclusively established as AI-generated. The most sensible approach is to begin with the default, then choose Balanced or Loose if Strict filtering removes too much relevant work from searches and feeds.

Can users report unlabeled AI artwork on ArtStation?

ArtStation users can now report artwork they believe is AI-generated but lacks an appropriate label. ArtStation says those reports will help improve filtering across the platform. The reporting option gives viewers a way to flag a possible mismatch between a project’s declared process and how the work appears to have been made.

User reports are not the same as a definitive finding that a creator used generative AI. A report adds a signal to ArtStation’s filtering system, and the platform’s Strict, Balanced, and Loose settings determine how strongly suspected content is filtered. This distinction matters because a reporting system can improve coverage of unlabeled work while still producing uncertain outcomes.

Users should report artwork only when they have a reasonable basis for concern and should avoid using reports to settle artistic disagreements. Creative work can include unusual styles, procedural tools, photo compositing, 3D rendering, and extensive digital edits that may look unfamiliar without being generative AI. The practical action is to use the reporting tool for suspected disclosure problems, not as a substitute for evidence.

Can the ArtStation AI filter reliably identify every AI image?

ArtStation’s AI filter cannot guarantee that every AI-generated image will be hidden or that every filtered image was made with generative AI. The platform’s suspected-content controls rely on signals that vary by setting, while creator declarations depend on accurate disclosure. The system is therefore useful for reducing unwanted content, but it is not an authorship-verification tool.

Strict filtering deliberately accepts a greater chance of filtering uncertain material because it activates when only a few signals suggest AI generation. That tradeoff can help viewers who prefer the broadest possible exclusion of suspected AI work. It can also create false positives, where human-made work receives filtering treatment because the available signals are incomplete or misleading.

Artist reaction to the new controls has been mixed, with concerns that strict user-driven filtering could both miss unlabeled AI work and incorrectly suppress human-made work. Independent artist-response coverage from Creative Bloq highlights that tension. For viewers, the practical conclusion is to treat a filtered feed as a preference-based browsing environment rather than a complete record of how every image was created.

What does the NoAI label mean on ArtStation?

ArtStation’s NoAI label is intended to tell AI systems that the marked content must not be used by them. Projects tagged NoAI receive an HTML NoAI meta tag, which ArtStation describes as an explicit prohibition on AI-system use of that content. The tag communicates a creator’s preference, but it does not independently prevent every outside system from accessing or using an image.

ArtStation also says it does not sell or license user artwork to other companies. ArtStation’s AI-content policy documentation states that portfolios should contain work users created and that creators should be transparent about their process. Those policies address how the platform handles user work, while the NoAI tag expresses a creator’s instruction to AI systems.

Creators should use NoAI when they want their preference stated in the project’s metadata, but they should not assume a metadata instruction provides complete enforcement outside ArtStation. Artists who need stronger control should also consider what they publish publicly, retain original files and process records, and document their work where appropriate. Creative professionals using automated design features should also understand how AI creative agents affect production workflows before applying broad labels to a project.

How should artists and viewers use the new ArtStation controls?

ArtStation creators should disclose generative-AI use accurately, while viewers should select filters that match their browsing preference. The platform’s 3 declaration categories and 3 suspected-content settings give both groups more specific options than a single all-or-nothing AI switch. Clear labels also help portfolio visitors understand whether an artist’s work reflects a generative, mixed, or non-generative process.

Viewers who want the strongest possible reduction in suspected AI work should keep the default Strict setting. Viewers who want to see more mixed or uncertain work can permit mixed workflows and choose Balanced or Loose suspected-AI filtering. ArtStation’s tools are designed for discovery preferences, so users should adjust them based on the kind of work they want to find rather than assuming one setting is universally correct.

Artists should preserve process evidence when their work may be questioned, including source files, sketches, project timelines, and layered documents where available. That record can be useful when a portfolio relies on a distinctive style or a complex digital workflow. The practical response is to use transparent labels, avoid unsupported accusations, and remember that platform filtering cannot resolve every authorship dispute.

FAQ

Is ArtStation hiding AI art by default?

Yes, ArtStation’s default configuration hides projects declared AI-generated, projects tagged CreatedwithAI, and projects associated with generative-AI tools. The default setting also applies Strict filtering to suspected AI content.

What is Strict filtering on ArtStation?

ArtStation Strict filtering hides suspected AI content when only a few signals suggest AI generation. Strict filtering removes more potentially AI-generated work, but it can also filter work that was not made with generative AI.

Can artists still post AI-generated work on ArtStation?

Yes, ArtStation does not ban AI-made artwork outright. ArtStation instead asks creators to disclose whether a project is primarily AI-generated, mixed, or 100% non-generative.

Can users report AI art that is not labeled?

Yes, ArtStation users can report artwork they believe is AI-generated but unlabeled. ArtStation says reports will help improve the platform’s filtering, but a report alone does not prove how a project was made.

Does ArtStation’s NoAI tag stop AI companies from using artwork?

No, ArtStation’s NoAI tag expresses a creator’s instruction through an HTML meta tag, but it does not guarantee that every outside AI system will comply. Creators should treat the tag as a stated preference rather than complete technical enforcement.