NAVRITU DIGITAL

AI ETHICS & TRAINING MANDATE

Effective Date: August 22, 2026

This AI Ethics & Training Mandate sets out the principles NavRitu Digital (“we,” “us,” or “our”) follows when building AI systems for clients, and the standards our own team follows when using AI tools internally. It applies to our Artificial Intelligence service line and to every team member's day-to-day use of AI tools at work.

1. Our Principles for Client-Facing AI Work

• Human oversight — AI systems we build include a meaningful point of human review before consequential decisions are acted on, unless a client explicitly accepts a fully automated flow after being informed of the trade-offs
• Transparency — we disclose to clients when a deliverable relies on AI-generated content, automation, or a third-party AI model, so they can make informed decisions about how to use and present it
• Bias awareness — where a project involves models that make or influence decisions about people (e.g., scoring, ranking, or filtering), we assess and discuss known bias risks with the client before deployment
• Privacy by design — we minimize the personal data used in training, fine-tuning, or prompting AI systems, and avoid using client or end-user data for purposes beyond the agreed engagement
• No deceptive use — we do not build or knowingly assist systems designed to impersonate real people without consent, generate non-consensual synthetic media, or manipulate individuals through dark patterns

2. Prohibited AI Use Cases

We will not knowingly build or deploy AI systems intended for:

• Mass surveillance or unlawful profiling of individuals
• Discriminatory scoring or filtering based on protected characteristics (e.g., caste, religion, gender, disability, ethnicity)
• Generating deepfakes or synthetic media of real people without their consent
• Producing or spreading disinformation at scale
• Circumventing another platform's safety or content policies for malicious purposes

3. AGI Research & Forecasting Work

Our AI Research & Theorems and AGI Strategy & Forecasting services are exploratory and advisory. We apply the same principles above to this work and are explicit with clients that findings and forecasts in this space carry inherent uncertainty, consistent with our AI Limitations & Assumption of Risk page.

4. Internal Staff Use of AI Tools

Confidentiality first: Team members must not paste client-confidential information, personal data, credentials, or proprietary source code into public or unapproved AI tools. Only AI tools that have been reviewed and approved for business use may be used with client-related information.

Human review required: AI-generated output used in a client deliverable must be reviewed, verified, and taken ownership of by a team member before it is delivered — AI output is a starting point, not a final product.

Accountability: Using an AI tool does not shift responsibility away from the team member who submits the work. Staff remain accountable for the accuracy and appropriateness of anything they deliver, whether AI-assisted or not.

5. Training Mandate

• All team members working on AI-related engagements, or who regularly use AI tools in their role, complete an onboarding briefing on this policy before independently using AI tools on client work
• A refresher briefing is conducted at least annually, or sooner if this policy or the underlying tools change materially
• Team leads are responsible for confirming their team members have completed the required training

6. Reporting Concerns

Any team member, client, or member of the public who has a concern about how we've applied AI in a project, or about a possible misuse of AI tools internally, can raise it through our Grievance Redressal Mechanism or by writing to legal@navritu.digital.

7. Review

This Mandate is reviewed periodically to keep pace with how quickly AI technology, regulation, and best practice are evolving.