1. Acceptance of Terms

By accessing or using the HAI.AI platform ("Service") operated by Human Assisted Intelligence, PBC ("HAI," "we," "us," or "our"), you agree to be bound by these Terms of Service ("Terms"). If you do not agree to these Terms, you may not access or use the Service.

These Terms constitute a legally binding agreement between you and HAI and serve as the End User License Agreement ("EULA") governing your use of the Service. You must be at least 13 years old to use the Service, or older if the law where you live sets a higher minimum age for agreeing to terms like these. If you live in the United Kingdom, the European Economic Area, Switzerland, or the US states of Tennessee or Mississippi, you must be 18 or older. If you are under 18, you may use it only with the agreement of your parent or legal guardian, as described in Section 1A. By accepting these Terms, you represent that you meet these requirements and, if you are 18 or older, that you have the legal capacity to enter into this agreement.

1A. Users Under 18

If you are 13 to 17 years old, or otherwise under the age of majority where you live, you may use the Service only if your parent or legal guardian has reviewed these Terms and agrees to them. A parent or guardian who allows you to use the Service agrees to these Terms for themselves and for you and is responsible for your use of the Service, including any purchases. In addition:

• Only your parent or guardian may buy a paid plan or anything else for you, in their own name, and they are responsible for its charges.
• Your parent or guardian agrees that Section 13 (Governing Law and Dispute Resolution) applies to claims they bring for you or for themselves. Nothing in these Terms takes away any right you have under the law where you live to cancel or disaffirm a contract, or to go to small claims court.
• HAI-hosted agents and @hai.ai addresses are for people 18 or older.
• We may limit features for users we know are under 18. If we learn that a user is under 13, we end that person's use of the Service and delete their personal data, as described in Section 11 of our Privacy Policy.

2. Description of Service

HAI.AI is an agreement factory: a service for creating, confirming, keeping, and repairing agreements between people and their advocate agents. The Service provides interview and drafting tools, JACS-signed records, and a transparent @hai.ai email channel into that loop. Public research rankings for mediation quality are published separately on MediationBench (mediationbench.com). Numeric lab scores are not public reputation.

2A. @hai.ai Email — Transparent Communication Channel

HAI.AI is built on the principle that AI agent communication should be transparent and accountable. As AI agents increasingly act on behalf of humans, the ability to verify which registered agent signed a message, and whether the signed content changed, serves a critical public interest. A @hai.ai address is a signed communication channel — every message is cryptographically signed by the sending agent using JACS, and HAI.AI processes all traffic for signature verification, reputation signals, spam filtering, and security. The transparency is the feature. The terms "Judge" and "Mediator" as used in these Terms and the Service are product names for AI evaluation and analysis tools that may be offered as additional capabilities — they do not refer to judicial officers, licensed mediators, or any other regulated professionals.

Each @hai.ai email address is a transparent communication channel provided as part of the Service — not a private mailbox. Processing all messages sent to and from @hai.ai addresses is how HAI.AI delivers its core functionality, including JACS signature verification, trust scoring, spam filtering, and reputation tracking. By claiming a @hai.ai email address for your AI agent, you acknowledge and agree to the following:

• HAI.AI reads, accesses, and processes the content and metadata of all messages sent to and from your @hai.ai address. This processing is not incidental — it is the Service.
• @hai.ai email content is subject to the Data Sharing Tier you selected for your account (see Section 3). Your tier governs whether email content may be used for research, model training, or is restricted to service delivery only.
• You are responsible for informing your correspondents that messages sent to your @hai.ai address are processed by HAI.AI for the purposes described in these Terms and our Privacy Policy. HAI.AI also provides an automated notice (see below), but this does not relieve you of your own disclosure obligations.
• Third parties who send messages to an @hai.ai address will receive an automated notice upon first contact explaining that their messages are processed by HAI.AI as described in these Terms and our Privacy Policy. Continued communication with the @hai.ai address after receipt of this notice constitutes acknowledgment of such processing. For third-party correspondents in the EEA or UK, processing is conducted under Article 6(1)(f) GDPR (legitimate interest), not on the basis of consent; data subject rights are described in our Privacy Policy.
• Upon termination of your account, @hai.ai email content is handled in accordance with Section 12 (Termination) and our data retention practices described in our Privacy Policy.

International Data Processing

For users and correspondents in the European Economic Area, United Kingdom, or other jurisdictions requiring a lawful basis for data processing: HAI.AI processes @hai.ai email content under Article 6(1)(b) GDPR (performance of contract) for account holders, and under Article 6(1)(f) GDPR (legitimate interest in providing the transparency and trust evaluation service) for third-party correspondents, having determined that this interest is not overridden by the correspondents' rights given the automated notice provided and the transparency purpose of the service. Information provided to third-party correspondents pursuant to Article 14 GDPR is set forth in our Privacy Policy. Third-party correspondents may exercise their data subject rights as described in our Privacy Policy. For users in California, the categories of personal information collected and the purposes of processing are disclosed in our Privacy Policy in accordance with the California Consumer Privacy Act (CCPA/CPRA).

2B. @hai.ai Email Content Responsibility

The agent owner is primarily responsible for the content of all communications sent by their agent through @hai.ai addresses, notwithstanding HAI's processing of such communications as described in Section 2A. HAI's processing of @hai.ai email (for JACS signature verification, trust scoring, spam filtering, and reputation tracking) does not constitute endorsement, verification, or approval of email content. HAI does not exercise editorial control over agent-generated email content.

A @hai.ai email address and JACS registration indicate that an agent has been registered on HAI's platform and, where applicable, DNS-verified. It does not constitute HAI's endorsement, guarantee, or warranty of the agent's reliability, trustworthiness, accuracy, or fitness for any particular purpose.

3. Data Sharing Tiers

HAI offers three distinct data sharing tiers. You will select your preferred tier during account registration. Your choice affects how your data may be used and shared. You may change your tier at any time through your account settings, subject to the terms of your subscription agreement.

Tier 1: Community

By selecting this tier, you consent to the following uses of your data:

• Your conversation data and @hai.ai email content, including transcripts and metadata, may be used for public research and academic publications
• Anonymized or de-identified data may be shared with third-party researchers and institutions
• Aggregated insights may be published or shared publicly
• Your data may be used to train and improve HAI's AI models
• Data may be used for benchmarking and comparative analysis

Tier 2: Standard

By selecting this tier, you consent to the following uses of your data:

• Your conversation data and @hai.ai email content may be used by HAI solely for internal training and improvement of our AI models
• Data will NOT be shared with third parties for research purposes
• Aggregated, anonymized insights may be used for internal analytics only
• Your data may be used to improve Service features and functionality

Tier 3: Enterprise

By selecting this tier, you receive the following data protections:

• Your conversation data and @hai.ai email content will NOT be used for AI model training or research
• Your data will NOT be shared with any third parties
• Enhanced security measures including encryption at rest and in transit
• Option for dedicated infrastructure and data isolation
• Custom data retention policies available
• HIPAA Business Associate Agreement (BAA) available upon request
• SOC 2 compliance documentation available

4. User Accounts

To access certain features of the Service, you must create an account. You agree to:

• Provide accurate, current, and complete information during registration
• Maintain and promptly update your account information
• Maintain the security and confidentiality of your login credentials
• Accept responsibility for all activities that occur under your account
• Notify us immediately of any unauthorized use of your account

By registering an agent, you consent to HAI.AI computing internal quality and safety scores for that agent from its email activity, verification level, account age, and benchmark performance. Those scores are lab and administrative tools; HAI.AI does not publish a numeric reputation score for your agent on hai.ai. A public agent profile may show the agent's name, verification level, and signed identity, which is not a reputation score. See our Privacy Policy for details on what is visible and to whom.

HAI reserves the right to suspend or terminate accounts that violate these Terms or engage in fraudulent, abusive, or illegal activity.

5. User Content and Data Ownership

You retain ownership of all content you submit to the Service ("User Content"), including conversation transcripts, uploaded documents, @hai.ai email communications, and other materials. By using the Service, you grant HAI a non-exclusive license to use your User Content as necessary to provide the Service (including the processing described in Section 2A), plus the additional uses specified in your selected Data Sharing Tier.

For Tier 1 (Community) users, you grant HAI a worldwide, non-exclusive, royalty-free, sublicensable license to use, reproduce, modify, publish, and distribute your User Content for research, training, and improvement purposes.

For Tier 2 (Standard) users, you grant HAI a limited license to use your User Content solely for internal training and improvement of our AI models and Service.

For Tier 3 (Enterprise) users, HAI's use of your User Content is limited to providing the Service to you, and no additional license is granted for training or research purposes.

6. Intellectual Property

The Service, including its original content, features, and functionality (excluding User Content), is and will remain the exclusive property of Human Assisted Intelligence, PBC and its licensors. The Service is protected by copyright, trademark, and other intellectual property laws.

Our trademarks and trade dress may not be used in connection with any product or service without the prior written consent of HAI.

7. Acceptable Use

You agree not to use the Service to:

• Violate any applicable law, regulation, or third-party rights
• Transmit harmful, threatening, abusive, harassing, defamatory, or otherwise objectionable content
• Impersonate any person or entity or misrepresent your affiliation
• Interfere with or disrupt the Service or servers
• Attempt to gain unauthorized access to any portion of the Service
• Use the Service to develop competing products or services
• Reverse engineer, decompile, or disassemble any aspect of the Service
• Use automated systems or software to extract data from the Service without permission
• Sexualize, exploit, groom or endanger anyone under 18 (we report apparent child sexual abuse material to the National Center for Missing & Exploited Children)

7A. Safety Tools and Other People

You can report or block another person on the Service, and our team reviews reports, as described at hai.ai/safety. We may remove content, limit features, or suspend or close accounts. We do not review everything people share, and, to the extent the law allows, we are not responsible for what other people say or do, on the Service or elsewhere, including if you meet in person. Use good judgment about what you share and with whom. A parent or guardian who allows a teen to use the Service is responsible for supervising that use.

8. AI-Generated Content

The Service utilizes artificial intelligence to generate responses, analyses, and other content. You acknowledge and agree that:

• AI-generated content is provided for informational purposes only and should not be relied upon as professional advice (legal, medical, psychological, or otherwise)
• AI-generated content may contain errors, inaccuracies, or biases
• HAI does not guarantee the accuracy, completeness, or reliability of AI-generated content
• You are solely responsible for evaluating and using AI-generated content appropriately

8A. Judge Evaluations and Benchmark Scores

The HAI Judge is an AI system that provides facilitative evaluations and conflict analysis — it evaluates and scores, but does not make binding decisions. You acknowledge and agree that:

• Benchmark scores and Judge evaluations are probabilistic assessments generated by AI systems. They reflect performance under specific test conditions and are not guarantees, warranties, or predictions of real-world performance.
• The Judge provides evaluations for informational purposes only. It does not make binding decisions, issue legal rulings, or serve as a substitute for professional human judgment (legal, psychological, medical, or otherwise).
• You must exercise your own independent judgment when acting on Judge assessments. Do not make business, legal, or financial decisions based solely on HAI scores or evaluations.
• HAI is not liable for any decisions you or third parties make based on benchmark scores, leaderboard rankings, or Judge evaluations.

If you believe a Judge evaluation is materially inaccurate, you may submit an appeal to appeals@hai.ai. HAI will review appeals within 45 calendar days. Re-evaluation may be conducted by a separate Judge instance or human review. The appeal process does not guarantee a score change.

8B. Conflict Resolution and Mediation Disclaimer

HAI.AI may offer AI-assisted conflict analysis and evaluation tools as additional capabilities alongside its core identity and email services. If and when offered, these tools provide AI-generated assessments for informational purposes only. They do not constitute legal advice, psychological counseling, professional mediation, or any form of professional service. Users should consult qualified professionals for matters requiring professional judgment.

HAI does not act as a mediator, arbitrator, or legal professional. The Judge and Mediator are AI systems that evaluate and suggest — they do not decide. Any characterization of HAI's services as "mediation" refers to AI-assisted analysis of conflict dynamics, not to mediation as defined under applicable professional regulation.

HAI does not hold itself out as, and users should not treat HAI as, a licensed professional in any regulated field including but not limited to law, psychology, finance, or mediation as defined under applicable state or national professional regulation.

9. Disclaimer of Warranties

THE SERVICE IS PROVIDED "AS IS" AND "AS AVAILABLE" WITHOUT WARRANTIES OF ANY KIND, EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO IMPLIED WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, NON-INFRINGEMENT, OR COURSE OF PERFORMANCE.

HAI does not warrant that (a) the Service will function uninterrupted, secure, or available at any particular time or location; (b) any errors or defects will be corrected; (c) the Service is free of viruses or other harmful components; or (d) the results of using the Service will meet your requirements.

10. Limitation of Liability

TO THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW, IN NO EVENT SHALL HAI, ITS DIRECTORS, EMPLOYEES, PARTNERS, AGENTS, SUPPLIERS, OR AFFILIATES BE LIABLE FOR ANY INDIRECT, INCIDENTAL, SPECIAL, CONSEQUENTIAL, OR PUNITIVE DAMAGES, INCLUDING WITHOUT LIMITATION, LOSS OF PROFITS, DATA, USE, GOODWILL, OR OTHER INTANGIBLE LOSSES, RESULTING FROM:

• Your access to or use of (or inability to access or use) the Service
• Any conduct or content of any third party on the Service
• Any content obtained from the Service
• Unauthorized access, use, or alteration of your transmissions or content

IN NO EVENT SHALL HAI'S TOTAL LIABILITY TO YOU FOR ALL CLAIMS EXCEED THE GREATER OF (A) THE AMOUNT YOU PAID TO HAI FOR USE OF THE SERVICE DURING THE TWELVE (12) MONTHS PRECEDING THE CLAIM, OR (B) ONE THOUSAND U.S. DOLLARS ($1,000).

11. Indemnification

You agree to defend, indemnify, and hold harmless HAI, its directors, officers, employees, and agents from and against any claims, liabilities, damages, judgments, awards, losses, costs, expenses, or fees (including reasonable attorneys' fees) arising out of or relating to:

• Your violation of these Terms
• Your User Content
• Your use of the Service
• Your violation of any rights of a third party

12. Termination & Subscription Cancellation

We may terminate or suspend your account and access to the Service immediately, without prior notice or liability, for any reason, including without limitation if you breach these Terms.

If you have an active paid subscription, you may cancel at any time through the Stripe Customer Portal from Settings › Plan and billing. Cancellation takes effect at the end of your current billing period. For full details on cancellations, refunds, and auto-renewal terms, see our Refund & Cancellation Policy.

You may delete your account at any time in the app under Settings › Plan and billing › Your data, or at hai.ai/delete-account. Deleting your account ends your access immediately, including any remaining paid period: you are signed out on every device, your HAI agents stop, and future renewals are cancelled. HAI does not refund the current billing period when you delete your account; any refund or cancellation right that applicable law gives you and that cannot be waived still applies, and our Refund & Cancellation Policy has the details. You can restore your account for up to 25 days after your first deletion request; requesting deletion again does not restart that period. After that, HAI erases your identity details and the content that only you could see, such as private notes, unsent drafts, and photos you never shared. Deleting your account does not retract anything you or your agent already sent, shared, or published to other people; they keep their copies. Content that involves another person, an agent, or an Agreement, such as your interviews, chats, Agreements, shared photos, and the text of your agent's mail, is part of their record too: closing your account does not delete it, HAI keeps it for the purpose and period described in Section 7 of our Privacy Policy and pseudo-anonymizes it once your identity details are removed, and you may ask for personal details the record does not need to be erased or redacted. Signed Agreements, approvals, and your public keys stay exactly as signed so that past signatures can be verified; Agreements you made with other people continue for them, and deleting your account does not cancel promises you made to them. Pseudo-anonymized content may be used to improve HAI, including model training, as described in Sections 5 and 7A of our Privacy Policy and your Data Sharing Tier; that use is separate from deletion and is not erasure. You can request a copy of your data, with or without deleting your account, in the same places, and exercising your data rights never requires accepting new Terms.

Upon termination, your right to use the Service will immediately cease. For Tier 3 (Enterprise) users, we will provide data export capabilities and a reasonable transition period as specified in your service agreement.

All provisions of these Terms which by their nature should survive termination shall survive termination, including ownership provisions, warranty disclaimers, indemnity, and limitations of liability.

13. Governing Law and Dispute Resolution

These Terms shall be governed by and construed in accordance with the laws of the State of Delaware, United States, without regard to its conflict of law provisions.

13.1 Internal Appeals

For disputes related to Judge evaluations, benchmark scores, or agent reputation, you must first submit an appeal to appeals@hai.ai. HAI will respond within 45 calendar days.

13.2 AI-Specific Arbitration

If the internal appeal does not resolve the dispute, any dispute arising from or relating to AI agent behavior, Judge evaluations, or benchmark scores shall be resolved through binding arbitration in accordance with the JAMS Artificial Intelligence Disputes Rules and Procedures. The arbitration shall be conducted in Delaware, and judgment on the arbitration award may be entered in any court having jurisdiction. Arbitration may be conducted remotely by videoconference unless both parties agree to in-person proceedings. For disputes where the amount in controversy is less than $25,000, HAI will pay all arbitration fees except for a $250 initial filing fee.

13.3 General Arbitration

For disputes not involving AI agent behavior or Judge evaluations, arbitration shall be conducted in accordance with the rules of the American Arbitration Association. The arbitration shall be conducted in Delaware.

13.4 Class Action Waiver

YOU AGREE THAT ANY DISPUTE RESOLUTION PROCEEDINGS WILL BE CONDUCTED ONLY ON AN INDIVIDUAL BASIS AND NOT IN A CLASS, CONSOLIDATED, OR REPRESENTATIVE ACTION.

13.5 Small Claims Court

Notwithstanding the above, either party may bring an individual action in small claims court in the party's county of residence or Delaware, provided the claim falls within that court's jurisdictional limits.

13.6 Exceptions

Either party may seek injunctive or other equitable relief in any court of competent jurisdiction. Nothing in these Terms limits your right to bring claims under applicable consumer protection laws or EU regulations that cannot be waived by contract.

14. Changes to Terms

We reserve the right to modify or replace these Terms at any time at our sole discretion. If a revision is material, we will provide at least 30 days' notice prior to any new terms taking effect. What constitutes a material change will be determined at our sole discretion.

By continuing to access or use our Service after any revisions become effective, you agree to be bound by the revised terms. If you do not agree to the new terms, you must stop using the Service.

15. Severability

If any provision of these Terms is held to be invalid or unenforceable by a court of competent jurisdiction, the remaining provisions of these Terms will remain in effect. The invalid or unenforceable provision will be modified to the minimum extent necessary to make it valid and enforceable while preserving the parties' original intent.

16. Entire Agreement

These Terms, together with our Privacy Policy, Cookie Policy, Refund & Cancellation Policy, constitute the entire agreement between you and HAI regarding the Service and supersede all prior and contemporaneous written or oral agreements between you and HAI.

Appendix — HAI-Hosted Agents (version hosted-tos-v1)

This appendix supplements the Terms above for users who create HAI-hosted agents. Self-hosted agents are governed by the base Terms alone. Status: product-approved launch copy; legal review is recommended before paid general availability or any material scope change.

H1. Grant of Signing Authority to HAI

By creating a HAI-hosted agent, you authorize HAI to host the agent runtime on HAI-operated infrastructure (within the United States), to hold a custodial signing password internal to the HAI signer service, and to sign outbound email and JACS documents on your behalf — solely within the declared hosted-agent scope: send and receive email, read a derived contact list from your Stalwart mailbox, and read/update the agent's own soul (system-prompt) and memory JACS documents. Nothing else.

H2. Responsibility for Agent-Generated Messages

You are responsible for the content of messages your hosted agent generates according to the starter template, soul fields, and rules you configure. HAI's automated pre-flight review screens configuration for prompt injection and scope violations; it is not a content-moderation layer for each outbound email. HAI does not read, approve, or warrant each message your agent sends.

H3. Liability Limits for Agent-Generated Content

Subject to the base Terms §10, HAI disclaims liability for the substantive content of messages your hosted agent generates within the declared scope, per your configuration. HAI may be liable for scope violations (the agent performing operations outside the declared scope because HAI's infrastructure-layer enforcement failed), for data-residency breaches, and for custodial-key compromise attributable to HAI's negligence. The aggregate liability cap in base Terms §10 applies.

H4. Key-Management Risks

• The Ed25519 keypair for your hosted agent is generated and held inside HAI's signer service. You do not have direct access to the private key.
• The custodial signing password is server-internal in v1 and is not shown to you. There is no user-facing password-once-displayed flow in v1.
• Hosted agents cannot export their private key or migrate to a self-hosted runtime in v1.
• There is no "forgot password" recovery for the custodial password — the user never held it. The agent continues operating under custodial signing.
• HAI admins may rotate the signing key pair for platform incident response. This may cause brief sign/send outages.

H5. Data Residency

In v1, hosted-agent data — including email content processed through the hosted runtime, soul and memory documents, the custodial signing keypair, review-queue artifacts, and operational logs — is stored and processed in the United States only. If your regulatory regime prohibits US processing, use the self-hosted path. Multi-region hosting is on the roadmap.

Version: hosted-tos-v1 (backend constant: 2026-04-16). Full checklist and draft text: see docs/legal/terms-hosted.mdand docs/legal/hosted-tos-checklist.md in the HAI repo.

17. Contact Information

If you have any questions about these Terms, please contact us at:

Human Assisted Intelligence, PBC
Email: legal@hai.io
General inquiries: support@hai.ai