Independent position
Comparable rating outputs that condense evidence into a common analytical language while preserving the drivers, cut-off date and evidence basis behind the conclusion.
Human Intelligence for AI
RATE AI LTD is incorporated in Ireland and built for a global operating and analytical field. RATE AI helps organisations use more AI with greater confidence by combining external, independent human judgement with decision traceability, dependency resilience, ratings and continuous intelligence.
The institution combines a published methodological foundation with proprietary intelligence infrastructure, evidence discipline and structural separation between methodology, data integrity, ratings and commercial work. Its committee and validation architecture is international by design.
What RATE AI is built to do
RATE AI builds and operates Human Intelligence for AI: independent human oversight around consequential AI operations, durable decision traceability, dependency resilience and an external rating and intelligence field that keeps those operations connected to the wider market.
Movement and materiality
Continuous intelligence across deployment, law, enforcement, incidents, dependencies and evidence changes that can move the previously observed position.
Defensible decisions
Decision Traceability designed to show what was known, how a decision was made, where human judgement entered and what can later be demonstrated.
See consequence early enough to preserve choice
Why we exist
AI systems can scale across institutions and borders at extraordinary speed. RATE AI exists to help organisations expand AI capability while preserving meaningful human judgement, decision traceability and the practical ability to intervene and change course.
What we aim to preserve
Decision autonomy: the practical capacity of institutions to understand, challenge, change, stop or defend consequential AI use while judgement, evidence and operating alternatives remain available.
Independence is structural
RATE AI separates methodology, evidentiary integrity, rating validation and procedural oversight across distinct functions, with independent checks governing the method, the evidence and the final rating.
Commercial mandates can provide human oversight, traceability, resilience work and deeper intelligence. Rating outcomes remain under separate independent governance, with new verified evidence entering the normal reassessment process.
Commercial engagement preserves independent rating authority. Clients may use RATE AI's specialist network, their own advisers or a mixed team while the rating process remains independently governed.
Separate responsibilities. One analytical standard
The governance model makes independence observable through distinct responsibilities, documented checks and structural separation.
Methodology Committee
Develops, calibrates and approves methodological revisions and scaling logic.
Authority: methodological revisions
Data Integrity Unit
Verifies evidentiary inputs, provenance, consistency and human-review status.
Authority: evidentiary integrity
Rating Committee
Reviews analytical outputs and validates band assignment under the frozen methodology.
Authority: rating validation
Ethics & Independence Panel
Oversees conflicts, conduct and procedural integrity and may suspend publication.
Authority: procedural integrity
Published method. Proprietary intelligence infrastructure
Transparency and intellectual property serve different purposes. RATE AI publishes enough methodology and evidence logic for institutional scrutiny. The data architecture that makes large-scale observation, comparison and monitoring possible remains a protected analytical asset.
The RATE AI Method
The methodological foundation explains how RATE AI reads exposure, governance capacity, evidence quality, rating bands and change over time. It is version-controlled so methodological evolution can be distinguished from real-world movement.
Evidence disciplineSource provenance, qualification and durable records.
Version controlMethodology and evidence cut-offs are preserved.
ValidationReliability, robustness and methodological challenge are part of the build.
Publication disciplinePublic conclusions remain separate from protected raw infrastructure.
Research and institutional network
RATE AI is designed to work with academic, technical, legal and sector expertise across jurisdictions where independent validation or specialist knowledge is required. The committee and validation architecture is structured for international membership and external challenge, so methodological scrutiny, rating judgement and sector expertise draw on more than one national ecosystem.
Data / IP boundary: RATE AI can disclose the existence and role of its Census and evidence libraries while keeping underlying datasets, taxonomies, source mappings, signal histories, entity-level comparative databases and internal query architecture proprietary.
Leadership
Raluca Antonie
Founder of RATE AI and builder of its methodological and institutional architecture, Raluca Antonie leads the development of an independent rating and intelligence institution designed to make consequential AI systems comparable, traceable and governable across jurisdictions, sectors and organizations.
Her work brings together legal and policy analysis, evidence-based methodology, public-interest governance and large-scale data architecture. RATE AI grows from this intersection: a system designed to rate AI and help institutions understand where exposure is accumulating, what decisions remain defensible and what changes materially over time.
Founder profile links to the independent professional and research record at ralucaantonie.com.
The book behind the method
THE RATE AI METHOD: Risk Intelligence. How AI Risk Is Detected, Rated, and Governed before It Becomes a Crisis provides the intellectual foundation from which the operational methodology and institutional architecture developed. On RATE AI, the book serves as a methodological foundation for the institution.
Read the Executive Impact Brief →External human judgement for consequential AI
Explore the public ratings and intelligence, or speak with RATE AI about an Independent Human Oversight mandate.
Independent intelligence across the market
RATE AI maintains a persistent external evidence and intelligence layer across systems, providers, peers, jurisdictions, enforcement, incidents and dependencies. Client mandates add private context to that wider field, allowing decision-makers to see both their own evidence and the environment moving around it.