iViRTH’s core work is structured data, identity and lifecycle records. Where analytics, simulation or AI-assisted features are used, they follow the principles below. We describe capabilities by status and do not present planned features as available.
1. Scope
These principles apply to analytics, indicators, simulation, predictive or optimisation workflows, and any machine-learning or AI-assisted feature iViRTH builds, integrates or uses in its services. Such features are offered only where implemented and demonstrated for the engagement.
2. Principles
- Transparency: we state what a feature does, its data sources, and its known limitations.
- Human oversight: outputs support decisions and do not replace accountable human review, especially where outcomes affect people, safety or compliance.
- Fitness for purpose: models and rules are used only with suitable data, and validated before being relied on.
- Accuracy and honesty: we do not overstate performance, and we label estimates, simulations and predictions as such.
- Privacy and security by design: minimise personal data and apply the safeguards in our security overview.
- Fairness: we look for bias and unintended effects where models affect people.
- Accountability: a named owner is responsible for each analytical feature in an engagement.
3. Use of client data
We do not use client data to train general-purpose models, or share it with other clients, without explicit written agreement. Where third-party AI services are used, we tell you, and use only providers whose terms prevent your data being used for their own training unless you agree.
4. Validation and monitoring
Analytical outputs are tested against agreed measures before use, and reviewed over time for drift, data-quality issues and errors. We document material assumptions and limitations.
5. Limits
AI and analytics can be wrong. Outputs are not legal, engineering, medical or regulatory advice, and should not be the sole basis for high-stakes decisions.
6. Using this website
This site does not use AI to make automated decisions about visitors, and does not profile visitors.
7. Raising a concern
If you think an analytical output is wrong, unfair or harmful, tell us and we will review it and, where needed, correct or withdraw it.
8. Contact
Email hello@ivirth.in with “Responsible AI” in the subject, or use the contact page. iViRTH, Uttar Pradesh, India.