Terms of Service | JudgeMyAI — AI Evaluation Engagement Terms
Legal · Master Terms

The terms behind the neural layer.

JudgeMyAI supplies elite human judgment to organizations that cannot afford misaligned AI. These Terms define that relationship — what we owe each other, who owns what, and where the lines are. Written to be read, not to hide behind.

Effective Aug 16, 2026 Version 3.1 Covers Site & Engagements Questions → legal@judgemyai.com
01

Acceptance of these Terms

By accessing judgemyai.com or engaging JudgeMyAI services, you agree to these Terms of Service. If you accept on behalf of an organization, you confirm you are authorized to bind it. If you do not agree, do not use the site or services.

Signed engagement documents (Statement of Work, Services Agreement, NDA, DPA) govern project-specific matters; these Terms remain the baseline wherever a signed document is silent, and the stricter protection always prevails.

02

What JudgeMyAI provides

JudgeMyAI delivers expert human intelligence for AI evaluation and alignment, including:

  • LLM response evaluation — claim-level grading of model outputs by credentialed domain specialists.
  • RLHF preference data — expert-ranked preference pairs and reward-model training data.
  • AI red-teaming — structured adversarial testing to surface jailbreaks and alignment failures.
  • Hallucination detection — verification of facts, citations, and reasoning chains.
  • Safety & compliance evaluation — assessment against medical, legal, and financial standards.

Talent-facing services (evaluator roles, calibration, cohort participation) are governed by the terms presented during application and the subsequent independent contractor agreement.

03

Engagements, scope, and delivery

3.1

Each engagement is defined by a Statement of Work specifying deliverables, quality thresholds (including inter-rater agreement targets), timeline, cohort composition, and price.

3.2

Standard deployment begins within 48 hours of signature from our pre-vetted expert bench; fully calibrated cohorts are typically production-ready within 7 to 14 days.

3.3

Deliverables that miss a contracted quality threshold are re-run at no additional cost or refunded pro rata, at the client's choice of remedy.

3.4

Pilot engagements of two to four weeks are available at standard rates and convert to continuous programs on request.

04

The covenant — mutual obligations

You Commit To
  • Supplying lawful data you have the right to share, free of third-party rights violations.
  • Accurate regulatory context (e.g., HIPAA, GDPR scope) for your domain before evaluation begins.
  • Using deliverables to improve AI safety — not to deceive or unlawfully harm end users.
  • Respecting our experts: no poaching outside the agreed engagement without consent.
We Commit To
  • Top-2% experts only — credential-verified, calibration-tested, never crowd workers.
  • Contracted quality thresholds, with re-run or refund when we miss them.
  • Confidentiality by default — NDA on first contact, per-client data isolation.
  • Honest limitations — we tell you what evaluation can and cannot certify.
05

Payment terms

Fees are stated in the Statement of Work. Unless otherwise agreed, invoices issue on delivery milestone and are payable net 30 days. Late amounts accrue interest at the lower of 1.5% per month or the maximum lawful rate. Clients may dispute invoiced items in good faith within 15 days of receipt; undisputed portions remain payable.

06

Intellectual property

  • Your data stays yours: prompts, model outputs, benchmarks, and internal materials you submit remain your property.
  • Deliverables are yours: evaluation reports, graded datasets, preference pairs, and red-team findings produced under an engagement are assigned to you on payment, except in open-license programs where release terms are stated in the SOW.
  • Our methods stay ours: evaluation methodologies, rubrics, calibration systems, and the vetting protocol remain JudgeMyAI's property. You receive a license to use them within the deliverables, not to resell them as evaluation services.
  • Site content: the website, brand, and published research are JudgeMyAI's work — cite freely with attribution, republish only with written permission.
07

Confidentiality

Both parties protect each other's confidential information with at least reasonable care, using it solely for the engagement. A mutual NDA can be executed at first contact, before any sensitive material is shared. Obligations survive termination for three years; trade secrets survive for as long as they remain trade secrets.

Client evaluation data is never used to train models, never published, and never shared between engagements. See our Privacy Policy for the data-handling detail.

08

Warranties and disclaimers

We warrant that services are performed with professional skill and care by qualified experts, and that deliverables meet the thresholds in your Statement of Work.

Apart from that, and to the maximum extent permitted by law, services are provided "as is". We do not warrant that evaluation guarantees model safety, that specific benchmark scores will improve, or that any deliverable makes a model fit for a particular regulated use. Human judgment dramatically reduces risk; it does not eliminate it. Deployment decisions remain yours.

09

Limitation of liability

Except for breach of confidentiality, IP infringement, or liabilities that cannot be limited by law, each party's aggregate liability is capped at the fees paid to JudgeMyAI in the 12 months preceding the claim. Neither party is liable for indirect, incidental, or consequential damages, including lost profits or model deployment losses.

10

Term and termination

  • Site terms apply while you use judgemyai.com; stop using the site to end them.
  • Engagements run for the SOW term. Either party may terminate for material breach unremedied after 15 days' written notice.
  • Convenience exits: continuous programs may be terminated with 30 days' notice; work performed to date is payable and delivered.
  • Survival: confidentiality, IP, liability caps, and payment obligations survive termination.
11

Governing law and disputes

Unless your signed agreement states otherwise, these Terms are governed by the laws of the State of Delaware, USA, excluding conflict-of-law rules. The parties will first attempt good-faith resolution for 30 days. Unresolved disputes proceed to binding arbitration under JAMS rules; small claims may go to court. Nothing here limits either party's right to seek injunctive relief for IP or confidentiality breaches.

12

General provisions

  • Changes: material updates are posted with a revised effective date at least 14 days before applying; continued use constitutes acceptance. Active engagements keep their signed terms.
  • Severability: if any clause is unenforceable, the remainder survives.
  • Entire agreement: these Terms plus your signed documents form the whole agreement on their subject.
  • Assignment: neither party assigns the agreement without consent, except to a successor in a merger or acquisition.
  • Contact: legal@judgemyai.com for legal notices; enterprise@judgemyai.com for everything else.
Core Competencies

The services these terms govern.

  • LLM Response EvaluationExpert-domain grading of large language model outputs for factual accuracy, reasoning quality, tone, and instruction adherence, performed by credentialed subject-matter specialists.
  • RLHF Preference Data CollectionHuman preference rankings, pairwise comparisons, and reward-model training data produced by top-2% experts for reinforcement learning from human feedback pipelines.
  • AI Red-TeamingSystematic adversarial testing of LLMs to identify jailbreaks, unsafe outputs, bias, and alignment failures before deployment.
  • Hallucination Detection & PreventionClaim-by-claim verification of model-generated facts, citations, and reasoning chains, with labeled failure data used for fine-tuning and guardrail construction.
  • AI Safety & Compliance EvaluationAssessment of model behavior against medical, legal, and financial safety standards, including regulated-industry documentation and audit support.
  • Domain-Specific Model TrainingCurriculum design and expert-led fine-tuning data for specialized fields including medicine, law, engineering, and scientific research.
Straight Answers

Frequently asked, honestly answered.

For evaluation engagements, yes — a Statement of Work or Services Agreement (with NDA and DPA where required) governs the project. These Terms govern use of judgemyai.com and apply as the baseline wherever the signed agreement is silent.

You own your model data, prompts, outputs, and the deliverables produced from them. JudgeMyAI retains ownership of its evaluation methodologies, rubrics, and calibration systems. Preference data ownership is defined per engagement — open-license programs release it publicly.

Engagements carry inter-rater agreement and accuracy thresholds in the Statement of Work. Deliverables that miss a contracted threshold are re-run at no cost or refunded pro rata — your choice of remedy.

Yes. Material changes are posted with a revised effective date at least 14 days before applying, and continued use after that date constitutes acceptance. Active engagements keep their signed terms.

Yes. Scoped pilots of two to four weeks are available at standard rates, with the option to convert to a continuous program and apply pilot learnings to the cohort plan.

Terms accepted. Judgment reserved.

You made it through the legal layer — the layer above it is 3,200 vetted experts waiting for your model. Or waiting for you.

Apply for AI Jobs

AI doesn't improve itself. Humans do. The ghost in the machine.