Document control
Responsible AI Policy And Principles
SecureXcel and its group companiesClassification: PublicVersion: 1.0 | Effective Date: 22-May-2026 | Owner: AI Governance Committee Review Cycle: Annual or upon significant regulatory/technology change
1. Purpose
SecureXcel and its group companies are committed to designing, developing, and operating Artificial Intelligence (AI) responsibly. This Policy publicly states our commitments to lawful, ethical, secure, and trustworthy AI across our products, platforms, and services. It complements our internal AI Management Policy and associated procedures, and reflects our adherence to applicable laws and regulations, including the Digital Personal Data Protection Act (India) and other relevant data-protection and sectoral requirements.
For the purposes of this Policy, “AI” includes Machine Learning (ML), Generative AI (GenAI), decision-support systems, and automation developed and operated by SecureXcel. The Policy sets out high-level principles for privacy-by-design, security, fairness, transparency, explainability, human oversight, reliability, and sustainability in the use of AI.
This Policy is aligned with the principles of ISO/IEC 42001 (AI Management System) and related SecureXcel governance standards. Alignment indicates conformance in spirit and structure and does not, by itself, constitute certification. Operational controls, thresholds, stage-gates, model registries, incident playbooks, and evidencing remain internal to SecureXcel and are shared with customers or authorities under appropriate confidentiality where required.
This is a public-facing Policy intended to provide transparency into SecureXcel’s approach to responsible AI. It is reviewed periodically and may be updated to reflect changes in technology, regulation, or business context. Where there is any inconsistency between this public Policy and SecureXcel’s internal standards and procedures, the internal documents prevail.
2. Scope
This Policy applies to AI capabilities used or provided by SecureXcel and its group companies, including Machine Learning (ML), Generative AI (GenAI), decision-support systems, chatbots, automation tools, and third-party AI services integrated into our offerings. It covers employees, contractors, and systems involved in AI life cycle.
3. Principles
SecureXcel commits to the following AI principles:
- 1. Lawfulness & Privacy-by-Design We comply with applicable laws (including India’s Digital Personal Data Protection Act and other data-protection regimes) and embed privacy, purpose limitation, data minimization, and user rights into AI design and operation as applicable in each country of operation.
- 2. Security & Resilience We protect AI systems and data with appropriate technical and organizational measures (e.g., access control, encryption, monitoring) and address adversarial threats, abuse, and misuse.
- 3. Fairness & Non-Discrimination We strive to detect and mitigate unfair bias through dataset curation, testing, and ongoing monitoring, and we apply guardrails and business rules to prevent harmful outcomes.
- 4. Transparency & Explainability We seek to provide appropriate explanations of AI-assisted outcomes and to disclose when users are interacting with AI-enabled features where feasible and appropriate.
- 5. Human Oversight & Accountability We maintain human review for sensitive or high-impact use cases and assign clear ownership for model decisions, risk treatment, and issue resolution.
- 6. Reliability & Safety We validate AI systems for fitness of purpose and monitor for drift, degradation, and emergent risks, updating, retraining, or decommissioning systems when required.
- 7. Sustainability & Responsible Use of Resources We aim to make efficient use of compute resources and consider environmental impacts where material.
- 8. External AI Engagements
We assess all external partners and service providers involved in AI for security, privacy, ethics, and compliance, and require appropriate documentation, contractual safeguards, and assurances to ensure Responsible AI standards are maintained.
4. Governance & Accountability
SecureXcel’s AI Governance Committee provides oversight for responsible AI, including policy stewardship, risk oversight, approvals for significant AI deployments, review of incidents and corrective actions, and continuous improvement. Roles and responsibilities for security, engineering, compliance, and business units are defined in internal governance documents and procedures (kept internal for security reasons).
5. Acceptable Use (Public Commitments)
- AI-assisted outputs used in our services are reviewed and validated commensurate with risk.
- When appropriate, AI-generated or AI-assisted content is attributed or disclosed to users.
- Employees and partners are trained on responsible AI and must follow SecureXcel’s internal controls.
6. Prohibited Use
The following are prohibited within SecureXcel’s operations and offerings:
- Creating or distributing content that is false, misleading, defamatory, or harmful (including deepfakes).
- Bypassing security or moderation controls or using AI for unauthorized access, exploitation, or unethical practices.
- Impersonation of employees, customers, executives, or any individual/entity for deceptive purposes.
- Deploying AI code/outputs without required security, privacy, and regulatory checks.
7. Model Lifecycle (High-Level Commitments)
- Design & Build: We apply secure development and ML-Ops practices, use approved data sources, and document model purpose, limitations, and data lineage.
- Validation: We test for accuracy, robustness, explainability, security, privacy, and fairness, with appropriate approvals before production use.
- Monitoring: We monitor accuracy, error rates, drift, stability, and fairness. If thresholds or conditions warrant, we retrain, roll back, or retire models.
- Retirement: We decommission models in a controlled manner and retain artifacts per our records policies.
8. Third-Party AI & Supply Chain
We evaluate third-party AI providers for legal, security, privacy, bias, hosting location, data handling, and incident obligations, and we require appropriate documentation (e.g., model cards or equivalent) and contractual controls. We maintain traceability and documentation for integrated third-party systems.
9. Risk Management & Impact Assessment
AI systems are subject to risk identification, assessment, treatment, and monitoring. Where appropriate, we conduct AI Impact Assessments (AIIA) to evaluate ethical, privacy, security, and societal considerations and to document mitigations before significant deployments.
10. Incident Reporting & Redress
We maintain procedures to detect, report, and resolve AI-related incidents (e.g., harmful outputs, bias breaches, data leakage). Users or stakeholders who wish to raise concerns about AI in SecureXcel services may contact our Compliance Team via the contact information provided on SecureXcel’s website. We investigate credible reports and take appropriate corrective actions, including notifications as required by law.
11. Training & Awareness
We provide periodic training on responsible AI practices for relevant personnel and specialized training for technical teams working with AI.
12. Review & Updates
This Policy is reviewed at least annually or upon significant regulatory or technological change. Updates will be published on our website with version history.
13. Disclaimer
This public Policy summarizes SecureXcel’s Responsible AI commitments. Operational procedures, thresholds, detailed controls, model registries, and playbooks are internal and shared with customers or authorities under appropriate confidentiality where required.