Proposed Amendment to the Code of Conduct: AI tools and Responsible Participation Policy

[OpCo]

The ArbitrumDAO forum is meant to provide a meritocratic space for DAO participants to contribute based on their best judgment. Such a space is underpinned by the lived context delegates bring, the thoughts and different viewpoints that help shape discussions, and, ultimately, the cohesive output of all stakeholders reaching consensus via vote. This is partly why the ArbitrumDAO has been such an active and high-signal community to date. AI tools can support that work, but they cannot replace it.

The use of AI tools to assist with drafting, translation, summarisation, research and other administrative tasks is welcome, however contributors remain responsible for the accuracy, relevance and originality of anything they publish.

Lately, the forum has seen a rise in unproductive or low-value contributions that appear to be largely or entirely machine-generated with little to no human review: content that is repetitive, generic, off-topic, restates existing discussion without adding new insight, or otherwise provides little substantive value.

This content is rarely malicious and may not meet the bar for spam, but in volume it dilutes discussion, buries substantive contributions, and makes it overwhelming for delegates and contributors to use the forum.

As such, we want to propose an amendment to the ArbitrumDAO Code of Conduct so that forum content remains productive, constructive, and genuine. This is also a good opportunity to remind all contributors that they should follow the community guidelines of the forum.

Proposed Amendment to the Code of Conduct

We propose the addition of the paragraphs below in the Code of Conduct.

AI tools and responsible participation in the ArbitrumDAO

The use of AI tools to automate administrative tasks, overcome language barriers, or process and summarize large blocks of information is expected. That said, the ArbitrumDAO discourages the irresponsible use of AI agents that may lead to the forum being littered with comments that do little to add value to the conversation.

Going forward, moderators of the forum (currently the Arbitrum Foundation and OpCo) will proactively monitor for comments that are off-topic or irrelevant to the main discussion, with the aim of mitigating the irresponsible use of AI agents. Moderators may use a variety of tools to assist their review, but moderation decisions will be based on the quality, relevance and intention of the contribution rather than solely on whether it appears to have been AI-generated. Repeat infractions will lead to suspension or, in severe cases, user deletion.

Moderators retain discretion to remove content that undermines productive governance discussions. Where appropriate, moderators may redirect off-topic discussions to a more suitable venue rather than removing them or message the author with similar suggestions. Enforcement will generally follow a graduated process:

1. Removal or redirection of the relevant content, with an explanation where practical.
2. A formal warning for repeated low-quality or disruptive contributions.
3. Temporary suspension for continued violations following a warning.
4. Longer suspensions or permanent removal from the forum in cases of persistent abuse, automated spam campaigns, or other serious or repeated violations.

Community members are encouraged to flag content they believe is off-topic, repetitive or otherwise inconsistent with these guidelines. Moderators will review flagged content on a case-by-case basis.

Procedure

This proposal will follow the review process as outlined in the Code of Conduct:

• It will be in the proposals category in the forum for 14 days.
• If a combined 5% of the DVP, calculated at the time of the forum post, raises objections in the comments below, the proposal will go to a Snapshot.
• If there are no objections, or the objections represent less than 5% of the DVP at the time of the forum post, the amendment will be optimistically approved.
• If approved, the OpCo will update the Code of Conduct to include the proposed text above.

[Arbit1]

While I agree with the concern to a certain extent, we need to be very careful about how this is moderated and enforced.

I use AI to help structure and clearly communicate my thoughts because English is not my first language. The ideas and opinions are mine; AI simply helps me express them more clearly and concisely.

I also want to be careful with how we define “repetitive” content. I have repeatedly raised concerns about the lack of utility for the ARB token because those concerns have not received a meaningful response from any Arbitrum-aligned entity. More recently, several delegates have raised similar concerns about the ARB token price, yet those concerns also remain largely unanswered. Maybe now it will start get some traction. A valid concern does not become low-value simply because it is repeated, especially when it remains unresolved.

I am concerned that this approach could create an avenue for censoring opinions that differ from those of the AAEs especially if individuals and delegates are trying to hold them accountable.

[MconnectDAO]

I broadly support this AI Code of Conduct, as spammy low‑signal AI content clearly harms Arbitrum’s governance discussions.
However, Arbitrum is a global ecosystem, and many contributors rely on AI to express their ideas clearly in a language that is not their first.

To avoid losing valuable contributors and ideas, I believe we need a few safeguards.

1. Explicitly recognise AI‑assisted posts used for language support as a positive use case, not inherently suspicious.

2. Define “low‑value” content clearly, so irrelevant spam is separated from repeated but reasoned governance concerns.

3. Introduce basic transparency and an appeal path for moderation, so dissenting yet valuable voices are not unintentionally suppressed.

4. Encourage high‑signal AI‑assisted research (data, risk analysis, multilingual explainers), as it directly improves DAO decision‑making quality.

5. Clarify that repeated, well‑reasoned concerns on unresolved core topics (like ARB utility or treasury design) are not treated as repetitive AI noise.

In short, we should reduce irresponsible AI spam without blocking the best AI‑assisted ideas and voices in the ecosystem. @Arbit1 @OpCo

[Arb_Junior]

strongly support this amendment.

We need to keep the forum high-signal and productive. Responsible Al use is welcome, but low-effort spam is hurting real discussion. Good move by OpCo.