Legasus AI Acceptable Use Policy

This policy is incorporated into the Legasus Master Subscription Agreement.

This AI Acceptable Use Policy (“AI AUP”) governs use of AI-Enabled Features provided by AIDAN LLC d/b/a Legasus. Capitalized terms not defined here have the meanings in the Master Subscription Agreement.

AI-Enabled Features are intended to assist qualified professionals and business users. They are not a substitute for appropriate human judgment, verification, supervision, or compliance with law and professional obligations.

1. Permitted Use

Subject to the Terms, Customer and Authorized Users may use AI-Enabled Features for lawful business and professional purposes, including:

drafting and revising documents;
summarization;
legal research support;
matter review and analysis;
document classification and extraction;
intake analysis and scoring;
workflow support;
preparation of communications;
calendar and task assistance;
mock-trial and litigation simulations;
financial or case-related calculations as decision support;
customer-specific firm knowledge and best-practice workflows;
client-service support; and
other uses Legasus makes available.

Permitted use remains subject to required human review and all applicable professional obligations.

2. Human Review and Professional Responsibility

Customer will not treat AI Output as authoritative merely because the Services generated it or provided a citation, confidence indicator, score, or recommendation.

Before AI Output is relied on, filed, signed, sent externally, used to advise a client, or used for a legal or similarly significant decision, Customer will provide meaningful review by a qualified person appropriate to the use case.

Customer remains responsible for:

factual verification;
citation verification;
legal research validation;
applicable-law analysis;
professional judgment;
privilege/confidentiality;
ethical compliance;
court rules;
deadlines;
client disclosures;
supervision of staff and technology; and
any final decision or action.

3. High-Impact and Legally Significant Decisions

Customer will not use AI-Enabled Features as the sole decision-maker, or as a substantial factor without meaningful qualified human review, for decisions that produce legal or similarly significant effects on an individual, including decisions involving:

access to legal representation;
employment;
housing;
credit or lending;
insurance;
healthcare;
education admissions;
government benefits;
essential services; or
another area regulated as high-impact automated decision-making.

Where law imposes disclosure, impact assessment, appeal, explanation, testing, documentation, or human-review requirements for automated decisions, Customer is responsible for compliance.

Legasus intake scoring is intended as decision support based on a firm's configured preferences and experience and does not itself accept or reject legal representation.

4. Evidence, Courts, and Legal Proceedings

Customer will not use AI-Enabled Features to knowingly:

fabricate evidence;
create or submit false affidavits, declarations, testimony, exhibits, citations, quotations, authorities, or factual records;
conceal that evidence was altered where disclosure is legally required;
impersonate a judge, juror, witness, attorney, client, or other person for the purpose of deceiving a tribunal or third party;
create deceptive synthetic media intended to be represented as authentic evidence; or
obstruct justice, tamper with evidence, or violate a court order.

AI-assisted drafting or simulation is permitted where lawful, appropriately reviewed, and not deceptively represented as authentic evidence or a real person's statement.

5. Privacy, Confidentiality, and Sensitive Information

Customer may submit confidential or sensitive information to AI-Enabled Features only when Customer has authority to do so and the applicable feature is authorized for that category of data.

Customer should exercise particular care with:

privileged information;
sealed or protected material;
Protected Health Information;
information about minors;
government identifiers;
highly sensitive financial information;
biometric information;
criminal-history information; and
information subject to protective orders or contractual restrictions.

PHI may be submitted only through services/configurations designated as HIPAA-eligible and under an executed BAA where required.

6. Prohibited Illegal or Harmful Uses

Customer will not use AI-Enabled Features to:

violate or facilitate violation of applicable law;
facilitate fraud, theft, identity theft, extortion, or unlawful deception;
violate privacy, publicity, confidentiality, intellectual-property, or other rights;
unlawfully discriminate against protected classes;
unlawfully profile, rank, or classify individuals in a manner prohibited by applicable law;
exploit a person's vulnerability in a manner prohibited by law;
facilitate unlawful harassment, threats, stalking, or abuse;
generate or distribute malware, credential theft, destructive code, or instructions designed primarily to compromise systems;
evade sanctions, export controls, or legal restrictions;
facilitate unauthorized practice of law or another regulated profession; or
use the Services in a way that creates a material risk of harm to persons, property, legal rights, or the Services.

7. No Model Extraction, Cloning, or Competitive Training

Customer will not, and will not permit another person to:

use AI-Enabled Features or AI Output to train, fine-tune, distill, develop, or improve a competing artificial-intelligence or machine-learning model, system, product, or service;
systematically extract AI Output to create a competitive training or evaluation dataset designed to replicate Legasus;
probe or extract system prompts, hidden instructions, weights, embeddings, orchestration logic, proprietary retrieval methods, or other nonpublic components;
use automated means to reproduce material functionality of Legasus;
bypass rate limits, entitlements, model protections, safety controls, or security measures;
reverse engineer AI-Enabled Features except to the limited extent a restriction is prohibited by law; or
misrepresent Legasus AI technology as Customer's own underlying model technology.

This Section does not prohibit ordinary internal security, legal, compliance, or procurement evaluation conducted in good faith and not for competitive replication.

8. Benchmarking and Publication

Customer may internally evaluate AI-Enabled Features for quality and suitability. Customer may not publish a benchmark, model-comparison result, vulnerability result, or technical analysis that discloses Legasus Confidential Information, security-sensitive information, nonpublic architecture, or material designed to enable replication without Legasus's prior written consent.

Nothing prohibits truthful statements required by law or protected by nonwaivable legal rights.

9. Voice, Recording, and Synthetic Media

Where AI-Enabled Features create, process, record, transcribe, or synthesize voice/audio, Customer is responsible for:

providing legally required notices;
obtaining legally required consent;
complying with recording and wiretap laws;
obtaining rights to use voices or recordings;
disclosing synthetic or AI-generated content where required; and
preventing deceptive impersonation or fraud.

Customer may not use voice functionality to impersonate another person without lawful authority or to create materially deceptive communications.

10. Client-Facing and Third-Party AI

If Customer exposes AI-Enabled Features or AI Output to clients, prospective clients, Client Users, or other third parties, Customer is responsible for:

deciding whether such deployment is appropriate;
configuring the experience;
supervising use;
providing legally required AI disclosures;
explaining material limitations where required;
providing privacy disclosures;
providing meaningful human review or escalation where required; and
not representing that Legasus independently provides legal advice to the third party.

11. Research and Citations

Customer must independently verify legal authorities, quotations, links, citations, statutes, regulations, cases, docket information, and research supplied or summarized by AI-Enabled Features.

A citation appearing in AI Output does not mean Legasus has independently validated that the authority exists, remains good law, supports the stated proposition, applies in the relevant jurisdiction, or is quoted accurately.

12. Mock Trial, Simulation, and Prediction

AI simulations of judges, jurors, opposing counsel, witnesses, panels, mediators, decision-makers, or other persons are hypothetical tools. Customer will not represent a simulated result as an actual person's opinion or as a reliable prediction of a real outcome.

13. Financial, Settlement, and Case Calculations

AI-generated or automated calculations involving damages, medical expenses, liens, settlement ranges, fees, costs, net recovery, interest, valuation, or financial projections are decision-support tools only and must be independently verified.

14. Experimental Reduced-Approval Functionality

Legasus may permit an Administrator, upon request, to reduce or bypass default human-approval controls for certain agentic actions. This functionality is experimental and not recommended for material legal, financial, client-facing, or irreversible actions.

If Customer requests such functionality, Customer:

knowingly assumes responsibility for actions taken under the configured authorization;
must establish appropriate scopes, permissions, limits, and monitoring;
remains responsible for all resulting communications, changes, filings, transactions, and third-party interactions;
must maintain legally required human review even if the product permits technical automation; and
understands Legasus may disable the functionality for security, compliance, or risk reasons.

15. Customer-Specific AI Learning

Customer may use Customer-Specific AI Configuration only with data and materials Customer is authorized to use for that purpose.

Customer will not intentionally use Customer-Specific AI Configuration to ingest or learn from:

unlawfully obtained material;
another firm's confidential information without authorization;
content Customer is prohibited from using for model or automated learning under contract or law; or
third-party proprietary material where the intended use violates applicable license terms.

16. Security Testing

AI prompt injection testing, adversarial testing, or safety evaluation is permitted only when conducted on Customer's own authorized environment in a manner that does not access another Customer's data, disrupt the Services, bypass access controls, or violate the Responsible Disclosure Policy.

AI Acceptable Use Policy