Ethical Use of AI in the ONIC Product Vetting & Certification System.
ONIC uses an AI-powered review system to support its mission of bringing trustworthy, science-backed dietary supplements to consumers. This page explains how that AI system works, who built it, what safeguards exist, and how ONIC ensures that technology augments human judgment without replacing it.
Every certification decision is made by qualified human assessors. The AI Engine helps these experts work faster and deeper - it does not make decisions for them. No product receives or loses ONIC certification based on AI output alone.
A curated AI Engine, not a general-purpose chatbot
ONIC's AI-powered review system is a curated AI Engine purpose-built for the nutraceutical and dietary supplement industry. Created and maintained by 20+ subject-matter experts including Phytochemists, PharmDs, medical doctors, food scientists, and data scientists, with dataset updates published monthly.
Neural networks for predicting ingredient metabolic pathways and pharmacological interactions
ML models for assessing ingredient synergies and potential contraindications
Cross-references formulations against country-specific approved ingredient lists and dosage limits
NLP-based analysis to extract efficacy endpoints, study quality indicators, and evidence strength
RAG Foundation: Validated datasets underpin every output - the system retrieves from curated data, not from open-ended text generation
AI operates within ONIC exclusively through ONIC Commercial Services (OCS), the commercial division. It plays no direct role in the independent certification decisions made by ONIC Certification Authority (OCA).
| Function | What AI Does (OCS) | What Humans Do (OCA) |
|---|---|---|
| Product Validation | Screens ingredient safety, flags contraindications, checks dosage ranges | Independent verification through physical testing, document review, expert judgment |
| Claim Verification | Cross-references claims against clinical trials, rates evidence strength | Score Domain D2 & D4 using the 20-point RNS rubric with human review |
| Scientific Rigor | Evaluates clinical evidence quality, maps evidence hierarchy | Minimum 2 human assessors review every decision. 100% human verification |
| STOF (Safety, Traceability, Origin, and Finished-product verification) Assessment | Models Safety, Traceability, Origin, and Finished-product verification dimensions | Full 5-domain scoring including on-site audits and independent lab verification |
Checking ingredients against 1M+ clinical papers, reducing pre-screening time by ~50%, applying the same framework to every product, every time.
Manufacturing realities, nuanced regulatory interpretation, consumer safety intuition, and contextual judgment that AI cannot replicate.
How ONIC prevents the AI Engine from compromising certification independence
Operates the AI Engine for formulation support, manufacturer matching, and regulatory guidance. Commercial side.
Independent product certification against the ONIC Responsible Nutrition Standard. Governed by Institutional Review Board (IRB).
Formulation data is NOT used for AI training without explicit, standalone written consent. Withdraw anytime.
AI-generated formulations are assigned to the startup upon delivery. OCS retains zero rights to your product.
Comprehensive Data Governance Protocol covering classification, access controls, breach notification, and cross-border transfers. GDPR (for international data subjects) is additionally observed as a compliance framework.
ONIC's AI Credibility Requirements are aligned with ISO/IEC 42001:2023 (AI Management Systems) and ISO/IEC 42006 (audit and certification of AI). No "black box" recommendations - every AI output includes human-readable explanations.
Immediate product hold, independent AI audit, regulatory notification
Program-wide review, re-calibration, enhanced quarterly monitoring
Data breach protocol per governance framework, immediate containment
Annual audits detect systematic bias in AI recommendations. Audit findings are reported to the Institutional Review Board (IRB), and corrective action is mandatory where bias is identified.
Safety, Assurance, Footprint, and Evidence
AI: Screens against toxicological databases, heavy metal limits, drug-nutrient interaction matrices
RNS: Safety Verification - Non-negotiable safety floor
AI: Validates label claims against clinical trial evidence, flags unsupported marketing language
RNS: Label Transparency - Claim substantiation requirements
AI: Traces ingredient origin, verifies supply chain, maps sustainability indicators
RNS: STOF Integrated Validation - Sourcing, GHG, traceability
AI: Cross-references 1M+ clinical papers, scores evidence quality, identifies gaps
RNS: Scientific Substantiation - Evidence hierarchy scoring
Yes, you can opt out entirely. Every applicant can request a fully human review with zero AI involvement.
No, AI won't improve your chances. The firewall ensures OCA has no knowledge of whether you used AI. Think of AI as a practice exam, not an answer key.
No. ECM facility certification uses on-site inspections by OCA assessors. EAL certification uses blind sample competency testing. Neither process uses AI as a decision input. AI is limited to OCS commercial services only.
No. The ONIC Certification Mark means qualified human experts independently verified the product against rigorous standards. Every ONIC-certified product is verified by at least two qualified human assessors.
| Document | Section | Coverage |
|---|---|---|
| Master Contract Framework | Governance Structural Firewall, IP, Data Governance | AI system separation, IP ownership, training data consent, breach notification |
| Master Document | AI Credibility Requirements | Advisory-only principle, human verification, transparency, ISO/IEC 42001/42006 |
| Governance Framework | Additional Governance | Data interoperability, annual AI audit, Independent Board oversight |
| Crisis Management Plan | AI Failure Response | Formulation error response, bias detection, incident protocol |
| Services Catalogue | AI Services | AI Engine access tiers, formulation support, IP assignment |
| Responsible Nutrition Standard | 5-Domain RNS | Human-assessed standard; AI feeds into but doesn't determine scoring |
AI Management Systems
Audit & Certification of AI Management
Product Certification Bodies (NABCB)
India's Data Protection Law
Information Security Management (ISO 27001 information security posture)
Experience the power of AI-driven nutrition validation and certification
Get Started