AI Policy

Effective Date: July 1, 2026
Last Updated: July 1, 2026
Version: en-US 2026-07-01
Contact: [email protected]

This AI Policy governs use of Txaion artificial intelligence, machine learning, voice, speech, likeness, digital twin, memory, retrieval, agentic, automation, and synthetic media features ("AI Features"). This policy is incorporated into the Terms of Service and works together with the Acceptable Use Policy, Privacy Policy, and Data Processing Addendum.

1. Core Principles

Txaion's AI Features are intended to support lawful, transparent, consensual, and human-centered uses of AI. Users must use AI Features in ways that respect privacy, safety, intellectual property, human dignity, consent, and applicable law.

2. User Responsibility

You are responsible for:

Inputs, prompts, files, voice samples, likeness materials, memory data, and other Customer Content you provide
Reviewing and validating AI outputs before relying on them
Obtaining all required rights, licenses, notices, consents, and permissions
Complying with laws that apply to your use case, jurisdiction, users, and content
Disclosing AI-generated or synthetic content when required by law or when a reasonable person could be misled
Ensuring your end users and administrators follow this policy

3. AI Output Limitations

AI output may be inaccurate, incomplete, outdated, biased, offensive, non-unique, or unsuitable for your purpose. AI output does not constitute professional advice, including legal, medical, mental health, financial, tax, investment, engineering, cybersecurity, or other regulated advice.

You must use qualified human review for important decisions, regulated contexts, public distribution, or any use that could affect rights, safety, finances, health, legal status, reputation, employment, education, housing, credit, insurance, or access to essential services.

You may use voice cloning, voice synthesis, speech imitation, likeness simulation, personality simulation, memory replication, avatar, or digital twin features only if:

You are the person being replicated or simulated, or you have verifiable authorization from that person
You have obtained all legally required written, verifiable, or otherwise valid consents for voice, likeness, biometric, privacy, publicity, and data protection rights
You have authority to use the source materials
You disclose synthetic or AI-generated nature where required or where omission would be misleading
You honor revocation, deletion, and withdrawal requests where required by law or contract

For deceased persons, minors, employees, contractors, public figures, or people in sensitive contexts, you must obtain authorization from the legally appropriate person or rights holder and comply with additional applicable law.

You must maintain reasonable records of consents, authorizations, source materials, revocations, retention periods, and deletion requests for voice, likeness, biometric-related, personality simulation, memory, avatar, and digital twin features for as long as required by law and for a reasonable period needed to address disputes, audits, or rights requests. Txaion may request evidence of authorization, require additional verification, restrict sharing or generation, or suspend the relevant feature if consent is missing, disputed, withdrawn, or legally insufficient.

If a person withdraws authorization or requests deletion where required by law or contract, you must stop using that person's voice, likeness, biometric-related data, private data, and digital twin materials and delete or disable affected outputs where required by law and reasonably feasible through the Services.

Additional consent, authorization, retention, deletion, and withdrawal terms may apply under the Biometric, Voice, Likeness, and Digital Twin Consent Notice when that notice is presented or required for a feature.

5. Prohibited AI Uses

You may not use AI Features to:

Impersonate, deceive, defraud, manipulate, or mislead any person or organization
Create unauthorized deepfakes, voice clones, avatars, personality replicas, or digital twins
Generate non-consensual intimate imagery or sexual content involving minors
Exploit, groom, endanger, or harm children
Harass, threaten, dox, extort, stalk, or intimidate others
Promote hatred, unlawful discrimination, violence, terrorism, or self-harm
Create malware, phishing, credential theft, cyber abuse, or evasion tools
Conduct unlawful surveillance, biometric identification, emotion recognition, or sensitive profiling
Make or materially automate high-impact decisions without required human review and legal safeguards
Generate deceptive political, civic, financial, employment, healthcare, legal, or consumer content
Forge evidence, records, signatures, identification documents, consent records, or authentication materials
Bypass safety systems, content filters, rate limits, access controls, or payment systems
Train, fine-tune, benchmark, or improve competing models or services without written authorization

6. High-Impact Decisions

AI Features may not be used as the sole basis for decisions involving employment, education, housing, credit, insurance, healthcare, legal rights, immigration, criminal justice, public benefits, biometric identification, critical infrastructure, emergency response, or other high-impact contexts unless Txaion has agreed in writing and you implement all required safeguards.

Required safeguards may include human review, bias and accuracy testing, documentation, user notices, appeal processes, risk assessments, monitoring, security controls, and compliance with applicable AI, privacy, consumer protection, employment, and civil rights laws.

7. Synthetic Media Disclosure

When AI output includes synthetic audio, image, video, text, likeness, voice, or interactive persona content, you must provide clear disclosure if:

The content is public-facing or commercial
The content could reasonably be mistaken for a real person, real event, real endorsement, or human-generated communication
The content involves a public figure, private individual, employee, customer, patient, student, voter, investor, or vulnerable person
Disclosure is required by law, platform policy, contract, or industry rules

Disclosure should be prominent, understandable, and appropriate to the medium, such as "AI-generated," "synthetic voice," "digital twin," or similar wording.

You must not remove, obscure, or misrepresent watermarks, metadata, provenance signals, or disclosures applied by Txaion or required by law, platform rules, or contract. You are responsible for ensuring that synthetic media is not presented as a real endorsement, real event, human communication, professional advice, official record, or identity verification unless that presentation is accurate and lawful.

8. Data Use and Training

Txaion processes AI Feature data as described in the Privacy Policy and DPA.

Unless Customer expressly authorizes otherwise in writing or through a product setting made available for that purpose, Txaion does not use Customer Content to train, fine-tune, or improve third-party foundation models or Txaion general-purpose AI models. Txaion may use aggregated, de-identified, anonymized, safety, telemetry, evaluation, and feedback data to improve reliability, security, abuse prevention, and product quality where permitted by law and contract.

Customer Content may be processed for inference, retrieval, transcription, speech synthesis, output generation, safety review, abuse prevention, support, debugging, logging, security, and service delivery. Txaion may route Customer Content to subprocessors listed in the Subprocessor List when needed to provide enabled AI Features.

9. Human Review and Moderation

Txaion may use automated and human review to detect abuse, enforce policies, improve safety, troubleshoot issues, respond to support requests, comply with law, and protect users and the public. Txaion may remove content, block outputs, restrict features, suspend accounts, or report unlawful conduct when appropriate.

10. Intellectual Property

As between you and Txaion, you retain rights in Customer Content and outputs subject to the Terms. You are responsible for ensuring that your inputs and outputs do not infringe third-party rights. Txaion does not guarantee that outputs are unique, copyrightable, trademark-clear, or non-infringing.

You may not use Txaion AI Features to intentionally copy, imitate, or exploit protected works, trademarks, voices, likenesses, styles, or identities in a way that violates law or third-party rights.

11. Safety Testing and Abuse Prevention

Txaion may limit prompts, outputs, uploads, file types, voice features, generation volume, API access, sharing, downloads, or account activity to reduce abuse. Txaion may require additional verification, proof of consent, or enterprise review for sensitive AI, voice, likeness, or high-impact features.

12. Customer Products Built on Txaion

If you build an application, bot, agent, workflow, integration, digital twin, or customer-facing product using Txaion, you must:

Provide accurate notices and terms to your users
Obtain consent and lawful bases for data processing
Implement safety, security, rate limiting, logging, and human escalation appropriate to risk
Avoid presenting AI as a licensed professional or real human unless accurate and lawful
Provide a way for users to report abuse or request privacy rights
Comply with Txaion policies and applicable laws
Flow down materially equivalent use restrictions, disclosure obligations, and consent requirements to your end users where your product exposes Txaion AI Features to them
Maintain records reasonably sufficient to demonstrate compliance with consent, disclosure, synthetic media, and high-impact decision requirements where legally required

13. Reporting AI Abuse

Report suspected AI abuse, impersonation, unauthorized voice cloning, deepfakes, harmful outputs, or policy violations to [email protected]. Include URLs, account identifiers, output IDs, screenshots, timestamps, and a short explanation where available.

14. Changes

Txaion may update this AI Policy as laws, technology, risks, and product capabilities evolve. Continued use of AI Features after an update becomes effective means you accept the updated policy.