AI + Compliance Glossary

The world's first AI Management System standard (2023). Voluntary, becoming a procurement gate for EU/UK/US enterprise deals.

India's Digital Personal Data Protection Act. Governs consent, purpose limitation, data-fiduciary duties and breach reporting for personal data.

The international standard for Information Security Management Systems. The security baseline enterprise buyers expect from an AI vendor.

Privacy Information Management System extension to ISO 27001. Bridges an ISMS to GDPR + DPDP data-privacy duties.

US audit report by AICPA covering security, availability, processing integrity, confidentiality and privacy. Type II covers 6–12 months.

EU General Data Protection Regulation. Sets the global bar for personal-data handling; DPDP borrows heavily from its clause structure.

US NIST AI Risk Management Framework. Voluntary, function-based (Govern/Map/Measure/Manage) approach to trustworthy AI.

US Health Insurance Portability and Accountability Act. Required for AI vendors handling US Protected Health Information (PHI).

Per-system evaluation of an AI deployment's risk to people. Mandatory under ISO 42001, expected under DPDP for high-risk processing.

DPDP classification for entities processing large volumes or sensitive personal data. Triggers extra duties: DPO, DPIA, audit.

Structured evaluation of a processing activity's privacy risk. Mandatory for SDFs under DPDP and for high-risk processing under GDPR.

AI architecture where a router activates only a subset of expert sub-networks per token. Delivers large capacity at lower inference cost.

Open standard by Anthropic for exposing tools, data and prompts to LLMs. NeuroCortex v3 supports 25+ MCP connectors at ₹1/call.

AI systems that plan multi-step actions, call tools and iterate autonomously. NeuroCortex v3's Agent Mode wraps this cleanly.

The token budget an LLM can attend to at once. NeuroCortex v3 supports a long-context tier with a ₹1 surcharge past the threshold.

The act of running an already-trained AI model against new input to get a prediction or generation.

Pattern where the LLM retrieves relevant documents from a vector store before generating an answer. Standard for enterprise Q&A.

Deep-learning model trained on large text corpora for language understanding + generation. NeuroCortex v3 is India's frontier LLM.

Attack where hostile instructions inside untrusted content override the intended system prompt. Top OWASP LLM risk (LLM01).

OWASP's top 10 security risks for LLM applications — prompt injection, data leakage, supply chain, model DoS, etc.

AI models, weights, hosting and audit accountability owned within a single jurisdiction. NeuroCortex v3 is India's sovereign AI.

NeuroCortex v3 mode that turns the chat into an autonomous multi-step agent with tool-use and web access.

NeuroCortex v3 reasoning mode that runs a hidden chain-of-thought and only shows the final polished answer.

Code-tuned sibling of NeuroCortex v3. Scaffolds and runs React+Vite apps in-browser via WebContainer. ₹5 per run.

Browser-based Node.js runtime by StackBlitz that lets NeuroCortex CodeMaster run React+Vite apps entirely client-side.

Optimising for AI-answer surfaces like ChatGPT, Perplexity, Gemini and Google AI Overviews. Speakable + FAQPage + facts.json.

Indian payments gateway. Bharat NeuroTech's INR billing rail — wallet top-ups, audit purchases, consult bookings.

Applied consumer neuroscience — using attention, emotion, memory and decision science to design pricing, packaging and creative.

The category Bharat NeuroTech coined — using software (not brain implants) to decode people, brands and AI systems.