Is an AI Arbitrator Fair? How DecidePlease™ Keeps AI Adjudication Honest

It is a reasonable question to ask before you trust any system with your dispute: is an AI arbitrator fair?

DecidePlease™ does not ask you to take fairness on faith. It is built so fairness can be checked — in the record both sides share, in rules that apply the same way to everyone, and in awards that must explain themselves. It is also honest about where a human is the better choice.

Both Sides See the Identical Record

Fairness starts with information. Every case lives in a private access-code room, and everything filed is visible to both parties at all times — every submission and every evidence item.

There are no private caucuses and no side conversations with the adjudicator. Neither party can make a point the other side never sees. The arbitrator console is separate — neither party sees the arbitrator's queue or notes — but the decision must be based on the shared written record.

If the AI reads it, you both wrote it, and you both can read it too.

The Same Rules Apply to Both Sides

Before submissions open, both parties must affirm the arbitration agreement, recorded with timestamps. Each party also makes a separate choice about opting in to a binding award. One side cannot bind the other.

During the rounds — opening statements, rebuttals, an optional questions round, deliberation, and award — each party gets one primary submission per round. Evidence is submitted the same way by both sides: structured entries with a title, description, and optional link, with no file uploads.

In AI modes, patented screening technology (U.S. Patent Nos. 11,025,567 and 11,757,805) scores every submission the same way — tone and temperature, frequency, and intent divergence, against a threshold that adapts to the exchange. The patents cover real-time intent scoring and message remediation in dispute conversations, not adjudication itself. Outcomes are the same for everyone: delivered, held with a suggested softened rewrite, briefly delayed, escalated, or — only for threats of harm — rejected. Every score is audit-logged.

Symmetry does not guarantee a result you will like. It means the process does not quietly favor whoever filed first or wrote more aggressively.

Awards Must Give Reasons and Cite the Record

An unexplained decision is hard to trust, whoever makes it. The AI Adjudicator works from the written record only. It matches the two openings to separate agreed facts from contested points, groups the dispute into issues, and writes findings with reasons that cite the actual submissions.

That citation requirement is a discipline. A finding should trace back to something a party actually said or listed, so both sides can follow the reasoning. The AI does not investigate, browse, or invent facts to fill gaps.

Money awards carry an additional guardrail: they are strictly bounded by the parties' own figures — never more than the claimant asked, never less than the respondent conceded. With no figures stated, the award is declaratory.

Advisory by Default Means Nothing Compels Without Consent

Every award is non-binding and advisory by default, and clearly bannered as such. An award becomes binding only when both parties explicitly opted in, with separate checkboxes, timestamped, before submissions open. Posture is derived from those recorded opt-ins only.

An advisory award gives both parties a reasoned, neutral assessment they can use to settle, negotiate, or decide next steps — without either side being compelled by a process only one of them agreed to be bound by. How awards are treated in practice varies by state and by agreement.

A Human Review Tier Exists

DecidePlease™ does not force an AI decision on disputes that need a human. There are three modes:

In `ai` mode, the AI Adjudicator reviews the written record and issues the award.

In `ai_assisted` mode, a human arbitrator leads. AI screens submissions and drafts the award for the arbitrator to edit and issue — the human decides.

In `human` mode, a human arbitrator handles the case with no AI screening at all. Submissions file exactly as written.

If your dispute turns on judgment calls a human should make, you can choose a human from the start rather than arguing with an AI outcome afterward.

The Honest Limits

AI adjudication works from text only. It cannot watch a witness hesitate, hear tone of voice in live testimony, or weigh demeanor. Complex credibility calls — deciding who to believe when accounts directly conflict — are a human's strength.

It is also only as complete as the record you build. A clear, well-supported submission helps; a vague one leaves any decision-maker with less to work from.

DecidePlease™ is built for disputes that can be fairly decided on a written record, and it gives you a human path for those that cannot.

If fairness means seeing everything the decision-maker sees, getting reasons tied to the record, and never being bound without clear consent, DecidePlease™ was designed to meet that standard. Start a private room, review the agreement together, choose the mode that fits — including a human — and proceed on a record both sides can verify.

FAQ

Can the AI favor one side because of how screening works? Screening applies the same scoring to submissions from both parties, and every score is audit-logged. Either side's submission can be held, delayed, or escalated under the same rules.

What if I do not trust an AI decision for my type of dispute? Choose `ai_assisted` or `human` mode instead. In `ai_assisted` mode a human arbitrator decides with AI support; in `human` mode there is no AI screening at all.

Is an AI award automatically binding? No. Awards are advisory by default. An award is binding only if both parties separately opted in, timestamped, before submissions opened.

General information, not legal advice. Related: how-ai-adjudication-works, human-arbitrator-option, why-binding-arbitration-requires-both-sides-to-opt-in