AI Ethics Charter

Responsible AI,

by design.

Responsible AI,

by design.

Our commitment to privacy, fairness, and transparency in AI dubbing and video translation — built into every layer of how Perso Dubbing works.

CAI member · Adobe

ISO/IEC 27001

KISA ISMS

GDPR compliant

CCPA compliant

PIPA compliant

Core Values

Integrity

Honesty and transparency in every policy decision and product choice.

Consent first

Explicit authorization required before any biometric data is used.

Accountability

Full ownership of our policies, decisions, and enforcement actions.

Transparency

Clear disclosure of how AI operates and how data is handled.

Privacy by default

Data minimization and automatic deletion as baseline behaviors.

Responsible innovation

Advancing AI capabilities while actively managing risk and harm.

Five Core Principles

01

Prevention of harmful use

Perso Dubbing is built for creative and professional use — not misuse. We enforce strict prohibitions backed by our Terms of Service.

Non-consensual deepfake content is strictly prohibited

No violent, hateful, or sexually explicit AI-generated material

Political campaigning or disinformation use is not permitted

Content infringing third-party rights or violating law is blocked

Terms of Service §22

02

Transparency & content provenance

As a member of the Content Authenticity Initiative (CAI), founded by Adobe, Perso Dubbing supports the C2PA standard. AI-generated content carries verifiable provenance metadata traceable to its origin.

C2PA-standard provenance metadata in every AI-generated output

Supports platform transparency and regulatory compliance

03

Fairness & bias mitigation

AI models trained on diverse demographics, languages, accents, and cultural backgrounds — designed to minimize systemic bias from the ground up.

Diverse training data across demographics and cultures

Human dignity respected across the full AI development lifecycle

Transparent operations so users can understand and trust outcomes

04

Privacy by design

Privacy is the default, not a feature. We never collect or use personal information without explicit consent.

Opt-in only — no data collection without explicit consent

Full transparency on what is collected and why, before you agree

GDPR (EU §34) · CCPA (California §35) · PIPA compliant

05

Consent & creator rights

Your voice, your content, your rights. Biometric data is encrypted and deleted the moment it's no longer needed.

Voice and facial data encrypted, stored separately from personal data

Biometric data deleted immediately upon service completion

Deletion requests honored at any time

AI-generated output from your content belongs to you

Terms of Service §4, §5

Enforcement & Review

Automated detection

Machine learning-powered scanning tools detect potentially non-permissible content at the point of upload and generation — before it reaches output.

Human oversight

Flagged content undergoes manual review by trained moderators to ensure fairness and accuracy. Automation and human judgment work together.

Action & Appeals

Enforcement

When a violation occurs

Perso Dubbing may remove content, block generation, suspend or terminate account access, or report to relevant authorities under Terms of Service §22–23.

Appeals

Contesting a decision

Users may appeal moderation decisions through our support channels with details of the flagged content. Appeals are reviewed independently where possible.

Evolution

Policies evolve

Our moderation policies evolve with emerging threats, user feedback, and regulatory changes — reflecting our ongoing commitment to responsible AI.

Security Monitoring

Data security

Encryption at rest & in transit

Biometric data stored separately

Auto-deletion on completion

ISO/IEC 27001 certified

KISA ISMS accredited

Application security

Regular security audits

Code review process

Vulnerability management

SSL/TLS enforced

MFA on all accounts

Organization

Incident response plan

Security training program

Acceptable use policy

GDPR · CCPA · PIPA frameworks

BCDR & disaster recovery

Frequently Asked Questions

What is Perso Dubbing's policy on deepfakes?

Perso Dubbing prohibits using its platform to create non-consensual deepfake content, generate violent, hateful, or sexually explicit material, or use the service for political purposes. These are enforceable violations under Terms of Service §22, under which the Company may suspend or terminate a Member's account at its discretion. Perso Dubbing is actively developing automated harmful content detection to reinforce these prohibitions at scale.

Does Perso Dubbing allow voice cloning of public figures?

No. Terms of Service §22 prohibits collecting, storing, or disclosing another person's personal information — including biometric voice data — without their consent, and prohibits misusing another person's identity or likeness. Generating content that uses a real person's voice without authorization is a prohibited act under §22 and may result in service suspension or account termination.

Does Perso Dubbing train its AI on my content?

Perso Dubbing may use uploaded content to improve service quality, as described in Terms of Service §4. Members may refuse consent to AI learning use, though this may restrict access to certain services. Enterprise customers should contact our team for specific data handling agreements.

How does Perso Dubbing label AI-generated content?

Perso Dubbing is a member of the Content Authenticity Initiative (CAI), founded by Adobe. Content produced on the platform supports the C2PA standard, which embeds verifiable provenance metadata into AI-generated output. This metadata identifies content as AI-generated and traces it to its origin, enabling transparent disclosure across distribution platforms and supporting regulatory compliance.

Is my voice data stored after dubbing is complete?

No. Under Terms of Service §5, biometric information — including voice and facial data — is collected solely for essential service provision, securely encrypted, and stored separately from other personal data. It is promptly destroyed after fulfilling its service purpose. Members may request deletion of their biometric data at any time. Perso Dubbing operates under GDPR (EU §34), CCPA (California §35), and Korea's Personal Information Protection Act (PIPA).

How does Perso Dubbing prevent AI bias?

Perso Dubbing trains its AI models on datasets representing a wide range of demographics, languages, accents, and cultural backgrounds to minimize systemic bias. Systems are designed to respect human dignity throughout the entire AI development lifecycle — from training data selection to output review. Perso Dubbing provides transparency about how its AI operates so users can understand and trust the outcomes.

Security certifications & compliance documentation

Full details on our infrastructure, ISO/IEC 27001, and compliance frameworks.

View Trust & Security