Automatically Link DMCA Takedowns to Repeat Copyright Infringer Enforcement
Avatar DMCA
Current Issue
VRChat's Copyright Policy currently states:
- Repeat Infringers.
VRChat's policy is (i) to respond expeditiously to valid, DMCA-compliant notifications of claimed copyright infringement by removing, or disabling access to, infringing material, and (ii) in appropriate circumstances, to terminate the accounts of users who are repeatedly or blatantly infringing copyrights.
However, in practice, there does not appear to be a direct and reliable connection between successful DMCA takedowns and repeat infringer enforcement.
When infringing content is removed following a valid DMCA takedown notice, copyright owners may still need to submit separate manual user reports through the Help Desk before account-level enforcement can occur. This creates unnecessary administrative work, increases the possibility of inconsistent enforcement, and allows repeat infringers to continue exploiting the gap between two separate workflows.
If such an integration already exists internally, it does not appear to be functioning as intended in practice.
Proposed Solution
Implement a direct and automated connection between the DMCA copyright workflow and the account enforcement system.
Automatic Copyright Strikes
Automatically issue a copyright strike to the account that uploaded the infringing content whenever content is removed following a valid DMCA takedown notice.
Automatic Repeat Infringer Enforcement
Once the predefined strike threshold is reached, automatically apply the appropriate account action under the Repeat Infringer Policy without requiring additional manual reports from the copyright owner.
A more integrated and automated workflow would help ensure the consistent implementation of VRChat's existing Repeat Infringer Policy and support the reasonable implementation of its copyright policy under 17 U.S.C. § 512(i).
Log In
ハッボホテル
自動化そのものではなく、どの段階を自動化するかを分けて考えるべきです。
通知受理からストライク計上までの自動化は、CCBill判決でCWIEの体系的な通知追跡が"reasonably implemented"と認められた運用の延長線上にあります。ここは問題視しません。
崩れているのは後段です。ストライクが規定回数に達した時点で、レビューや異議申立てを一切挟まずに垢措置が自動発火する設計になっています。この提案は二つの段階を区別せず一括りに自動化しようとしていますが、閾値到達時に何が起きるかについては何の記述もありません。
Automation isn't the issue. Which step gets automated is.
Notice-to-strike tallying is basically what CCBill already found "reasonably implemented" when it credited CWIE's systematic tracking. Nothing wrong with that step.
The second step is where this falls apart. Once the strike count crosses the threshold, account action fires with nobody looking at it and no way to object. The proposal treats both steps as one block instead of two, and says nothing about what happens to the people caught by the second one.
你亲爱的酒保K
[1/3]
*Machine translation is used for this post
I do not claim that my view is the only correct standard. Rather than trying to defeat this position at all costs, the discussion should focus on how copyright disputes can be resolved fairly for everyone involved.
// TLDR
Link DMCA cases to account records automatically, but do not turn removals into automatic strikes or bans.
Under VN3, Public visibility may be permitted, Private-only, or prohibited depending on the asset's actual terms. Where the upload and Private use are authorized, and Force Private effectively prevents the disputed public access, it should be the default interim response.
Full takedown needs evidence that the upload or Private use is unauthorized. Copyright holders and authorized agents may request account action with sufficient evidence, but a human reviewer should decide every strike, suspension, or ban.
// Body
DMCA cases should be connected to repeat-infringer enforcement. A copyright owner or authorized agent should not have to file a second ordinary report just to make an existing copyright case visible to the people who decide account action.
That does not justify a system where a DMCA removal automatically becomes a strike and a strike threshold automatically becomes a ban. A content decision and an account decision answer different questions. The first asks whether access to a particular item should stop now. The second asks whether the account holder has engaged in conduct serious enough to lose access to the platform.
The DMCA is a notice-and-takedown process for identified material. It allows a provider to remove material or disable access to it. It does not require automatic strikes, a fixed strike count, or automatic account termination. A notice can justify a fast content response and an investigation. It cannot, by itself, establish that a player should receive a cumulative account penalty.
An account can hold years of purchases, friendships, community ties, creator work, and personal history. A wrong content restriction can often be undone. A wrong strike, suspension, or ban can interrupt relationships and events, damage a player's standing in a community, and make the account less reliable even after it is restored. Account penalties therefore need a human decision supported by evidence.
你亲爱的酒保K
[2/3]
VN3 is particularly relevant here because it treats VRChat upload and third-party use as separate permissions. It is a configurable licence template, not a single rule that applies the same way to every asset. The selected permissions, special provisions, licence version, and any individual authorization determine what the user may do.
Under VN3:
- C covers uploading an avatar to a social communication platform, including VRChat, for the user's own use.
- E separately covers uploading an avatar so that third parties can use it on that platform.
For E, a creator may permit Public release, permit only limited or Private sharing, prohibit third-party use, or require individual permission. VN3 itself uses VRChat Public and Private as examples of different sharing scopes. A Public avatar therefore does not have one universal meaning under VN3.
Public visibility does not transfer ownership. It may give other players permission to use the avatar when the licence allows that scope of use. If the licence permits upload and Private use but limits third-party use to Private or another restricted scope, the issue is public availability. It is not necessarily the avatar's continued presence in the uploader's account.
In that situation, Force Private or an equivalent access restriction should be the immediate response, provided it effectively prevents the disputed public access. It can limit public cloning and broad third-party access while preserving any personal use that the licence still permits. A full takedown needs a separate basis: the upload itself is unauthorized, Private use is prohibited, the licence has been revoked, or a visibility restriction cannot stop the reported access.
Force Private is not lenient enforcement. If Public access is the problem, it addresses that access directly. Removing more than necessary without a separate reason can also harm the rights holder by cutting off legitimate purchasers from permitted Private use, creating avoidable disputes, and making the creator's licence look less predictable than it is.
你亲爱的酒保K
[3/3]
An automated system cannot safely settle the facts that matter here. It cannot determine from a DMCA notice and a Public status alone whether the report identifies the correct asset and licence, whether C and E allow the use, whether special provisions or individual permission apply, whether reports are duplicates, whether the claimant has the proper authority, whether a notice was withdrawn or countered, or whether the account was compromised.
If the system counts those unresolved questions as strikes, it treats ambiguity as evidence against the player. The player then has to disprove a penalty that the system imposed before anyone reviewed the full record. That is not an acceptable basis for serious account moderation.
Copyright holders and authorized DMCA proxy agents should still be free to request account action. They can submit the relevant licence terms, proof of authority, upload history, repeated Public violations, prior warnings, evidence of evasion, and evidence that the conduct was deliberate rather than a single visibility error. That evidence can support a strike, suspension, or ban after a prompt human review.
VRChat can automate case linkage, evidence retention, duplicate detection, repeat-upload detection, and escalation. It can also apply an immediate visibility restriction when Public access is the verified issue.
However, a full takedown should be used when the upload or Private use is unauthorized, or when a narrower restriction will not stop the violation. Strikes, suspensions, and bans should remain human decisions based on the licence, the evidence, the account history, and the seriousness of the conduct. The account holder should receive clear reasons and a practical way to correct an error.
- siri -
The suggested measures would be harmful to the VRChat ecosystem.
- They would allow for uncontrolled abuse of the already flawed DMCA system.
- To submit a counter-claim, you would have to give out all of your personal information, while all you see from the claimant is their name and email.
- They would effectively kill off the public avatar scene, which is fair use since it constitutes a derivative work.
- The person behind "AvatarDMCA" consistently misuses DMCA law by applying it to license enforcement. Making an avatar public does not automatically constitute copyright infringement. It is, at most, a license covenant violation.
I would suggest the following instead:
- Ban all of AvatarDMCA's accounts for systematic abuse of the VRChat API and harassment of platform's users.
- Add safeguards to DMCA claim submissions to reduce the number of frivolous claims.
- siri -
Automatic translation to Japanese:
日本語への自動翻訳:
提案されている対策は、VRChatのエコシステムに悪影響を及ぼすものです。
- これらの対策は、すでに欠陥のあるDMCA制度の野放しな悪用を許すことになります。
- カウンター請求を提出するには、自分の個人情報をすべて開示しなければならない一方で、請求者側については名前とメールアドレスしか分かりません。
- これらの対策は、公開アバターシーンを事実上壊滅させることになります。これはフェアユース(公正利用)に該当し、二次的著作物を構成するものです。
- 「AvatarDMCA」の背後にいる人物は、DMCA法をライセンス強制の手段として常習的に悪用しています。アバターを公開設定にすること自体は、著作権侵害には自動的に該当しません。それはせいぜいライセンス規約(コベナンツ)違反にすぎません。
代わりに、以下のような対策を提案します。
- VRChat APIの組織的な悪用およびプラットフォームユーザーへの嫌がらせを理由に、AvatarDMCAの全アカウントを禁止(BAN)する。
- 不当・軽率なDMCA請求の件数を減らすため、DMCA請求提出プロセスにセーフガードを追加する。
WubTheCaptain
- siri -
- #1: A valid concern.
The rest of the post falls apart.
- #2: You can use a legal representative and their contact details on your behalf for a counter-claim. (This isn't changed by the proposal in OP.)
- #3: Your copyright on a derivative work and the subsequent free use claim may be thin in many scenarios, there's no simple answer. (See: The VRChat Copyright Trap by Derafog for a detailed explanation.)
- #4: He has a legal right to claim copyright infringement - and imo it isn't misuse from what I've seen. (This isn't changed by the proposal in OP.)
As for the latter:
- #1: Using output logs to gather avatar IDs is a feature within VRChat, as is browsing publicly available avatars or putting them on an avatar pedestal. I don't see this to be in violation of any Community Guidelines. There's no evidence of harassment I've seen, please don't attack anyone. If you must, you can make a moderation request at https://vrch.at/moderation privately.
- #2: DMCA § 512(f) already exists. Service providers generally have no say in claims against its users and must follow the legal process to avoid becoming liable for user-generated content, so I don't see what safeguards you're asking for.
vral
WubTheCaptain There is been systematic harassment.
WubTheCaptain
vral You can submit a moderation request at https://vrch.at/moderation, it can't be discussed/investigated/actioned on a public forum like this.
ふあゆき
DMCAの核心は「Notice and Takedown」であり、形式さえ整っていれば違法性を実精査することなく、即座にアクセスを遮断することがプラットフォームの免責条件となると理解しています。
つまりは「有効なDMCA削除通知後のコンテンツ削除」と「そのコンテンツが真に誰かの権利を侵害している事」は根本的に異なり、等号で結べるようなものではないという訳です。
よって拙速な自動化は、悪意ある第三者の虚偽通報によって無実のユーザーが排除される致命的なリスクを生むだけでなく、被通報者に保障されるべき「異議申し立て」の権利や、その手続きが完了するまでの適正な猶予期間を奪うことになるため、プラットフォームの運用としては、おおよそ不適切と言わざるを得ません。
I understand that the essence of the DMCA lies in the "Notice and Takedown" mechanism, where the platform's safe harbor immunity is conditioned upon expeditiously blocking access to content as long as the notification meets the formal requirements, without any substantive review of its actual illegality. In other words, "the removal of content following a valid DMCA takedown notice" and "the fact that the content genuinely infringes upon someone's rights" are fundamentally different concepts that cannot simply be equated. Therefore, rushing to automate this process would not only create a fatal risk where innocent users are wrongfully banned due to malicious false reports by third parties, but it would also deprive the accused users of their guaranteed right to file a counter-notification and the proper grace period required to complete that process. As such, it must be said that such an implementation is highly inappropriate for platform operations.
Avatar DMCA
ふあゆき
I understand the concern, and this is an important point when discussing DMCA systems.
I also understand that a DMCA takedown action and a final determination of actual infringement are not necessarily the same thing.
This proposal is not intended to apply penalties based solely on a DMCA notice.
The purpose is to ensure that, after VRChat has completed a valid DMCA-based takedown action, the result is properly reflected in the existing Repeat Infringer Policy workflow.
Existing DMCA procedures and protections, including counter-notifications, would remain in place.
Regarding account enforcement, my understanding is that final decisions are made through an appropriate review process.
The intended automation is to trigger the Repeat Infringer Policy workflow once a predefined strike threshold is reached, rather than replacing human judgment.
This is not a proposal to create a new enforcement standard, but to improve the system so that the existing policy can be applied consistently.
Manual review may be sufficient when the number of cases is small.
However, as the volume increases, relying entirely on manual processes creates risks such as incomplete history tracking, inconsistent handling, and unnecessary burden on copyright owners.
The purpose of automation is not to make legal decisions automatically, but to ensure that cases requiring review and action are identified accurately and consistently.
追記：
> DMCAの核心は「Notice and Takedown」であり、形式さえ整っていれば違法性を実精査することなく、即座にアクセスを遮断することがプラットフォームの免責条件となると理解しています。
このご意見に書かれている免責条件とは、17 U.S.C. § 512(c)(1)(C)に定められているものであると思います。
DMCAには複数のセーフハーバー要件があり、今回の提案は17 U.S.C. § 512(i)(1)(A)に関するものとなります。
表現に曖昧な部分があり、意図が十分に伝わっていなかったかもしれません。
ハッボホテル
Avatar DMCA
提案本文の記述はこれと異なります。閾値到達時にVRChatは「automatically apply the appropriate account action」を行うべきだと明記されており、追加の手動報告は不要とされています。これは垢措置そのものが自動発火するという意味で、審査を待つワークフローではありません。
「my understanding is that final decisions are made through an appropriate review process」の根拠は、VRChatの公開されたCopyright Policy・ToS・Community Guidelines・サポート文書のどこにも見当たりません。その仕組みが実在してここに適用されるなら、名指しした上で提案文に明記すべきです。前提のまま済ませる話ではありません。
The proposal's own text says something different. It says that once the threshold is reached, VRChat should "automatically apply the appropriate account action," no additional manual report required. That's the account action firing automatically, not a workflow that waits on review.
"My understanding is that final decisions are made through an appropriate review process" isn't documented anywhere in VRChat's public Copyright Policy, ToS, Community Guidelines, or support docs. If that process exists and applies here, name it and put it in the proposal. Don't just assume it.
Avatar DMCA
ハッボホテル
This may be a translation issue.
It differs from what I originally intended.
I understand that while VRChat already has several automated systems in place, the final decision is made by the moderation team.
Since I have already presented my revised opinion, there is no need to discuss this point.
ふあゆき
Avatar DMCA
先にハッボホテルが書いてくれている通り、提案本文の記載は「DMCAによる複数回の削除」をトリガーとしてアカウントBANを実行すると読み取れます。
とりあえず投稿者の意図はそこに無く、あくまで既存のポリシーを一貫して適用するためのワークフロー改善にある点は理解しました。
しかしながら、そもそもアカウントBANはユーザーの資産を剥奪する重大な措置であり、そこには「人間の介在」という安全弁は不可欠であると考えます。
現状、DMCA削除後に権利者が別途報告を行わなければならない仕様になっているのは怠慢ではなく、「DMCA対応」と「ユーザーの権利剥奪」という性質の異なる二つの判断を慎重に切り離しているバッファであると解釈しています。
仮に自動化するにしても「DMCA削除＝即ストライク確定」ではなく、「DMCA削除＝ストライクの保留」とし、ユーザーに異議申し立ての機会を通知し、一定期間（例：14日間）に異議申し立てがない、あるいは「手動精査」にて異議が却下された場合に初めて「正式なストライク」としてカウントする等が適当かと思います。
As ハッボホテル pointed out, the wording in the original proposal can indeed be read as executing account BAN directly triggered by multiple DMCA-based removals. That said, I now understand that this was not the author's actual intention, and that the goal is simply to improve the workflow so that the existing policy can be applied consistently.
However, an account BAN is a severe disciplinary action that completely strips users of their assets, meaning that the safety net of "human intervention" is absolutely indispensable. I interpret the current specification—where copyright owners must submit a separate report after a DMCA takedown—not as administrative negligence, but as a deliberate buffer designed to carefully separate two fundamentally different judgments: "expeditious DMCA compliance" and "the deprivation of user rights."
Even if some form of automation were to be introduced, it should not operate under a rigid "DMCA removal = immediate strike confirmation" model. A more appropriate implementation would be to treat a DMCA removal as a "pending strike." The system would then notify the user of their right to file a counter-notification; a "formal strike" would only be counted if no counter-notification is received within a certain period (e.g., 14 days), or if the submitted counter-notification is officially rejected after undergoing a thorough "manual review."
Avatar DMCA
ふあゆき
Thank you for your thoughtful feedback.
I agree that account enforcement should continue to involve appropriate human review, and that the design of the review process is an important consideration.
However, my proposal is not intended to change VRChat's existing copyright policy or how the Repeat Infringer Policy is applied.
Those policy decisions are ultimately up to VRChat.
The purpose of this request is simply to improve the system that supports the existing workflow, so that completed DMCA actions can be tracked consistently and repeat infringement history does not depend on separate manual reports or manual record-keeping.
Whatever review process VRChat chooses to use—whether it includes manual review, a waiting period, counter-notifications, or other safeguards—could still operate on top of the same underlying system.
ｺﾛｲﾄﾞ
自動化を適用した場合、
悪意を持ってライセンス違反を行っている層に対して、DMCAを悪用してこっちのアカウントに対して被害を及ぼす簡易的な手段を与える事になるのではないかと思うので慎重になったほうがいいのではないかと思います
Automating this could give malicious license violators an easy way to weaponize the DMCA against our accounts as users. I think the abuse case deserves more discussion before this ships.
WubTheCaptain
ｺﾛｲﾄﾞ The DMCA provides a legal remedy for this under § 512(f).
ｺﾛｲﾄﾞ
To be clear about what's on the table: an automated ban issued on top of the DMCA takedown itself. The DMCA already provides a counter-notice process precisely because a claim is not a finding of infringement. Auto-banning short-circuits that — the accused user is punished before they ever get a chance to respond. That's a separate problem from the one your argument addresses.