Choosing the arbitrator is one of the most consequential decisions in any arbitration.
That is always true. In AI disputes, it becomes even more important because the case may involve technical evidence, sensitive records, shifting system behavior, and a lot of room for confusion disguised as confidence.
The right arbitrator is not simply the person who sounds the most “techy.”
The right arbitrator is the person who can manage the process intelligently, understand the evidence well enough to ask disciplined questions, and reach a defensible result without being overwhelmed by jargon or seduced by novelty.
Start with the real question
Do not ask only:
- “Who knows AI?”
Ask:
- Who can decide this particular dispute well?
That is a much better selection standard.
An AI case may involve:
- software and vendor issues,
- licensing and data-rights questions,
- confidentiality problems,
- training-data or provenance questions,
- employment or consumer sensitivity,
- or evidence disputes involving logs, prompts, and model outputs.
The neutral does not need to be the world’s most famous AI commentator.
The neutral does need the right mix of judgment, process skill, and subject-matter fit.
What matters most
1. Subject-matter fit
There is no single archetype called an “AI arbitrator.”
The relevant expertise depends on the case.
For example:
- a model licensing dispute may call for licensing and technology-contract experience,
- a training-data conflict may call for IP, data-use, and provenance literacy,
- a workplace AI dispute may call for employment sensitivity,
- a consumer-facing chatbot dispute may call for stronger fairness and disclosure instincts,
- and a cybersecurity-related AI dispute may call for security and incident-response fluency.
The first goal is not to find a neutral who knows everything about AI. It is to find one whose background fits the actual dispute profile.
2. Process discipline
Many AI cases go sideways because the process becomes sloppy before the merits are decided.
The arbitrator should be able to manage:
- technical submissions,
- phased evidence exchange,
- confidentiality protections,
- expert sequencing,
- and realistic timetables.
This is especially important where the dispute includes logs, code-adjacent material, evaluation records, or trade-secret-sensitive evidence.
3. Comfort with technical evidence
The neutral does not need to write models or engineer systems.
But the neutral should be able to handle evidence that includes:
- prompts,
- outputs,
- benchmark or evaluation materials,
- version histories,
- incident reports,
- safety testing,
- and expert explanations of system behavior.
A neutral who cannot distinguish between a technical question and a storytelling shortcut may have trouble in a serious AI case.
4. Confidentiality instincts
AI disputes often involve trade secrets, customer data, internal safety methods, system instructions, and proprietary documentation.
That makes confidentiality more than a boilerplate issue.
The arbitrator should understand the practical importance of:
- protective measures,
- limited access protocols,
- careful hearing management,
- and disciplined treatment of sensitive records.
This is one of the strongest reasons to think carefully about arbitrator selection early.
5. Disclosure discipline and independence
California’s disclosure standards are a useful reminder that independence is not only a state of mind. It is also a recordkeeping and relationship issue.
Standard 7 and Standard 9 underscore that arbitrators must disclose matters that could raise doubt about impartiality and must make reasonable efforts to inform themselves about what should be disclosed.
In AI-heavy fields, where the same experts, law firms, vendors, institutions, and repeat players may appear often, disclosure discipline matters a great deal.
That does not mean an arbitrator with sector experience is disqualified. It does mean the parties should value real disclosure rigor, not assume it.
6. Judgment about AI tool use
One subtle but increasingly important factor is how the arbitrator thinks about AI use inside the arbitration itself.
Questions to consider:
- Does the arbitrator understand the difference between supportive tool use and judgment substitution?
- Does the arbitrator have a careful approach to confidentiality and external tools?
- Does the arbitrator communicate clearly about disclosure and verification?
This is especially relevant in light of the AAA-ICDR and Ciarb guidance, as well as the growing visibility of AI-assisted and AI-led arbitration products.
7. Writing clarity
AI cases can become dense fast.
A strong arbitrator should be able to:
- isolate the real issue,
- write clearly about technical facts,
- avoid being intimidated by complexity theater,
- and produce a reasoned result that parties can understand.
That is not cosmetic. It affects legitimacy.
8. Temperament and case-management style
Some AI disputes need an active manager.
Some need a steady minimalist.
Some need someone skilled at narrowing issues early and keeping experts from taking over the whole case.
The right style depends on:
- how much discovery is expected,
- how technical the proof will be,
- whether the relationship is already broken,
- and whether the parties need a fast or deeply reasoned process.
Style is not secondary. It is part of fit.
A practical selection framework
When evaluating a candidate, ask:
- Does this person have experience with disputes like this one?
- Can they handle the kind of evidence this case will generate?
- Do they seem disciplined about confidentiality and disclosure?
- Can they manage experts and technical submissions well?
- Are they likely to understand the real risk without being distracted by buzzwords?
- Do they have the temperament this case needs?
Those questions are usually more valuable than a generic search for someone who “knows AI.”
Red flags
Be careful if:
- the candidate’s expertise is broad but not relevant,
- they appear dazzled by technical language rather than able to test it,
- they have weak disclosure habits,
- they seem casual about sensitive data handling,
- or their style is a poor fit for the complexity of the matter.
Another red flag is assuming a famous retired judge is automatically the best fit. Sometimes that is true. Sometimes the better choice is a neutral with stronger technology-contract, privacy, licensing, or process-specific experience.
How institutions can help
AAA and JAMS both emphasize the value of expertise in arbitrator selection.
AAA’s panel materials highlight industry-specific panels and curated matching. Its October 10, 2024 press release on AAAi Panelist Search also shows that institutions are investing in better ways to identify suitable neutrals.
That does not replace party judgment. But it means parties should use institutional tools and case-manager input strategically rather than treating selection as a formality.
FAQ
Should I choose a lawyer, technologist, or former judge?
It depends on the dispute. The better question is which background best fits the actual issues and process needs.
Does the arbitrator need deep technical credentials?
Not always. Many AI disputes require strong process judgment and evidence discipline more than deep engineering specialization. But the neutral should be comfortable with technical material.
How important is disclosure in AI disputes?
Very important. Repeat-player relationships, sector concentration, and institutional overlap make disclosure discipline especially valuable.
Should the arbitrator’s own AI-tool practices matter?
Yes. Tool use, confidentiality judgment, and disclosure approach can affect trust in the process.
What is the biggest selection mistake?
Choosing based on surface-level AI branding instead of dispute-specific fit.
Conclusion
The best arbitrator for an AI dispute is rarely the person with the most impressive abstract AI profile.
It is the person who fits the case: the evidence, the confidentiality demands, the relationship dynamics, the legal issues, and the process pressure. That is how selection should work in any serious arbitration. AI just makes the consequences of getting it wrong more obvious.
Further Reading
- AAA About Our Panels: https://www.adr.org/panel/about-our-panels/
- AAA Arbitration Services overview: https://www.adr.org/arbitration/
- AAAi Panelist Search press release, October 10, 2024: https://www.adr.org/press-releases/aaa-icdr-launches-new-aaai-panelist-search-to-enhance-panelist-selection-with-ai-technology/
- JAMS Arbitrator Selection: How to Select the Right Arbitrator for Your Case: https://www.jamsadr.com/insight/2014/arbitrator-selection-how-to-select-the-right-arbitrator-for-your-case
- California Standard 7. Disclosure: https://courts.ca.gov/cms/rules/index/ethics/ethics7
- California Standard 9. Arbitrators’ duty to inform themselves about matters to be disclosed: https://courts.ca.gov/cms/rules/index/ethics/ethics9
