The Creative Intelligence Index is an independent framework for evaluating frontier AI companies against the values that matter most to society — transparency, human dignity, accountability, and moral courage.
Explore the Index"An AI system's character is revealed not by its benchmarks, but by who it protects when no one is watching."
Frontier AI companies shape the conditions of human thought, labor, creativity, and social life. Yet their internal values — and how faithfully they embody them — remain largely opaque.
The Creative Intelligence Index is an independent, evidence-based evaluation that names what we believe matters and holds institutions publicly accountable to it. It is not a government body. It carries no legal authority. Its power is reputational, scholarly, and moral.
Developed and maintained outside the AI industry. No funding from companies under evaluation. Scorecards produced by a panel of ethicists, technologists, and policy scholars.
Every score is grounded in publicly available evidence: filings, leadership statements, product documentation, terms of service, published research, and legal proceedings.
Companies are notified of draft scores and given 30 days to provide corrections of fact before public release. Their responses are published alongside the scorecard.
Scorecards are updated annually. Companies can improve. The index is designed to reward genuine progress, not to freeze reputations in place.
Each domain isolates a distinct dimension of institutional character. Together they form a comprehensive portrait of whether an AI company's stated values are operational, or merely aspirational.
Does the company disclose what its systems can and cannot do? Does it correct public misunderstandings promptly and forthrightly?
Do product decisions prioritize user and societal wellbeing, or engagement and monetization? How are vulnerable users protected?
Who is responsible when harm occurs? Are governance structures independent? Is internal dissent protected or suppressed?
How does the company engage with questions of labor displacement, concentration of power, and democratic institutions?
Has the company taken costly principled stands? Does it prioritize doing the right thing when it conflicts with commercial interest?
The CII uses a structured, documented methodology so that scores can be understood, contested, and improved. Our goal is not a verdict — it is a conversation.
Researchers compile public-source evidence for each domain: SEC filings, published research, leadership interviews, product documentation, legal proceedings, and press coverage.
An independent panel of ethicists, technologists, and policy scholars applies the 0–10 rubric to each domain, with written rationale required for every score.
Draft scorecards are sent to companies for factual review. Corrections of fact are incorporated; substantive disagreements are published alongside the final score.
Final scorecards, methodology, evidence summaries, and company responses are published simultaneously on this site.
Each domain is scored 0–10 by the panel. Scores reflect the preponderance of available evidence, weighted for recency and significance.
The CII's five domains are grounded in ten foundational values. These are not abstract ideals — they are operational commitments that can be observed, evidenced, and measured in the conduct of AI institutions.
A commitment to accuracy, correction of error, and honest representation of what AI systems can and cannot do.
The inherent worth of every person as the foundational constraint on how AI systems are designed and deployed.
Clear lines of responsibility, accessible grievance mechanisms, and genuine consequences when harm occurs.
Decisions made at the most local appropriate level. Resistance to unnecessary centralization of power or influence.
Long-term orientation toward the common good, including future generations who cannot yet advocate for themselves.
Willingness to take costly principled stands, protect internal dissent, and prioritize ethics over commercial pressure.
Active concern for those most vulnerable to AI's disruptive effects — workers, children, and marginalized communities.
Protecting and amplifying the distinctive creative capacity of human beings, rather than displacing or commodifying it.
Proportionate caution in the face of genuine uncertainty about transformative and potentially irreversible technologies.
Fair distribution of AI's benefits and burdens across income, geography, and generation — not only to those who can pay.
The CII is not a closed institution. We are actively building a community of scholars, ethicists, policymakers, technologists, and engaged citizens who share the conviction that AI accountability requires an independent public voice.
Get in touchWe are recruiting ethicists, technologists, legal scholars, and policy experts to serve on the v0.1 scoring panel. Commitment: approximately 20 hours over three months.
Help compile and verify the public-source evidence base for each company evaluation. Ideal for graduate researchers and policy analysts.
Universities, think tanks, and civil society organizations are invited to co-publish scorecards and contribute to the methodology's ongoing development.
Add your name to the founding declaration: a public statement affirming that AI companies must be held to meaningful ethical standards by independent voices.