Our Commitment to Ethical AI Music in India

1. Zero Non-Consensual Voice Cloning

We strictly prohibit and actively filter any attempt to synthesize unauthorized voice replicas of living or deceased playback legends, classical maestros, or public figures.

When users train custom voice models on our Pro Studio tier, our audio ingestion pipeline verifies consent and requires acoustic liveness tokens. Private custom voices are locked to individual user accounts and never exposed to the public model directory.

2. Responsible Data Curation & Artist Compensation

Unlike unvetted automated web scrapers, DhunAI collaborates with independent Indian session vocalists, classical instrument masters (sarangi, sitar, bansuri, ghatam, dholak), and sound engineers.

• Our core vocal presets are trained on consenting studio sessions with fair compensation and licensing agreements.
• Instrumental sample libraries are recorded in professional studios across Mumbai, Chennai, Kolkata, and Jalandhar to capture authentic regional acoustics and playing nuances.
• We continually explore revenue-sharing models with classical performers to support traditional arts education.

3. Preserving Regional Languages & Dying Musical Forms

Many global music models focus exclusively on Western 4/4 pop music. This risks marginalizing India's rich regional micro-genres — from Bhojpuri sohar to Bengali baul, Garhwali folk, Maand, and Carnatic ragas with complex talas like Mishra Chapu or Khanda Gati.

DhunAI dedicates at least 40% of its research computing resources specifically to vernacular phonetics, tonal nuances, and regional instrument preservation, ensuring India's musical diversity thrives in the digital age.

4. Anti-Hate Speech & Cultural Respect Guardrails

Music has profound emotional and spiritual significance in India. Our platform implements multi-layer semantic and acoustic safety classifiers:

• Harmful hate speech, communal abuse, casteist slurs, and defamatory lyrics are intercepted before audio generation begins.
• Devotional and traditional prompts (Bhajans, Shabads, Qawwalis, Kirtans) are handled with reverent musical sensibilities, adhering to respectful lyrical structures.

5. Content Authenticity & Watermarking

To promote transparency in the media ecosystem, DhunAI embeds cryptographic, inaudible acoustic watermarks conforming to C2PA (Coalition for Content Provenance and Authenticity) guidelines into all generated WAV and MP3 containers. This allows platforms to identify AI-generated tracks and verify their provenance.

6. Copyright Takedown & Grievance Redressal (IT Rules, 2021)

In accordance with the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021, and the Copyright Act, 1957, any copyright owner or interested party may report infringement, impersonation, or platform abuse directly to our designated Grievance Officer: