What SaferAI does

SaferAI is a France-registered non-profit organization (association loi 1901, reg no. 919 185 199 00016) headquartered in Paris that develops quantitative risk models, conducts independent company evaluations, and shapes international AI safety standards. Its core technology is a portfolio of risk modeling frameworks adapted from high-reliability industries (aviation, nuclear, chemical processing, finance) that translate AI capability measurements into probability estimates of real-world harm. The organization has completed nine cyber risk models and is actively extending methodology into loss-of-control, CBRN threats, and harmful manipulation scenarios. SaferAI also maintains the public AI Company Risk Ratings Platform (ratings.safer-ai.org), which evaluates frontier AI developers across risk identification, severity assessment, safeguard implementation, and governance, and provides Frontier Safety Framework advisory services to companies seeking compliance with the EU AI Act, EU Code of Practice, and California SB53.

The organization operates primarily through donations and charitable contributions rather than commercial pricing, and serves three customer segments: frontier AI companies seeking risk management and regulatory compliance support; governments and national AI Safety Institutes seeking evidence-based governance methodologies; and international standards bodies (OECD, NIST, ISO, CEN-CENELEC) seeking technical input on AI risk management specifications. Its go-to-market combines independent public ratings that create market accountability, direct advisory engagements with AI developers, and active participation in EU Code of Practice working groups, the OECD Hiroshima AI Process, CEN-CENELEC JTC 21, and ISO standards committees. The 17-person team blends research scientists, policy leads, and a standards lead, with advisors including technical staff from Anthropic, former OECD and UK DSIT leadership, and the former head of strategic foresight at OECD.

Name

SaferAI

Legal name

SaferAI (registered non-profit, reg no. 919 185 199 00016)

Website

https://safer-ai.org

Company type

Private

Operating status

Operating

Headcount range

1–10 employees

Short description

SaferAI is a Paris-based non-profit that builds quantitative AI risk models, publishes independent ratings of frontier AI companies' risk management practices, and shapes international AI safety standards (EU AI Act, OECD, ISO) for governments, regulators, and frontier AI developers.

Ownership category

akta.pro rank

Product category

AI Risk Management

akta.pro primary industry

Regulatory Readiness & Audit Automation (e.g., EU AI Act, NIST AI RMF, ISO/IEC 42001) (HDAAAMAE)

akta.pro secondary industries

Responsible AI, AI Governance & Compliance Services (BPAEAHAJ), Audit, Explainability & Accountability Tooling (traceability, reporting) (HDAAAKAL), Enterprise AI Governance, Risk & Compliance Platforms (Model Risk, Audit, Policies) (HDAEANAE), Model Transparency, Explainability & Interpretability (XAI) (HDAAAMAB)

Keywords

• AI risk management
• AI safety research
• Quantitative risk modeling
• AI governance standards
• AI company ratings

Headquarters

HQ city

Paris

HQ country

France

HQ region

Europe

Offices1 record

Markets served

GTM type

B2B

Offering type

Services

Cost components

Personnel, Technology or R&D, Marketing or Sales, Operations

Revenue model

1. Donations and Charitable Contributions: As a registered French non-profit (association loi 1901), SaferAI operates primarily through donations and charitable contributions. The organization maintains scientific rigor while advising governments and evaluating companies, accepting donations to fund independent research.

Go-to-market motion2 records

Distribution channels3 records

Marketing channels4 records

Core offering

SaferAI is a French non-profit that develops quantitative AI risk models translating AI capabilities into measurable harm estimates, independently evaluates and rates leading AI companies on their risk management practices, and provides advisory services to help companies develop Frontier Safety Frameworks compliant with emerging regulations such as California's SB53 and the EU Code of Practice. The organization also leads AI risk management standards development in the EU and contributes to international standards bodies (CEN-CENELEC, ISO, OECD) to shape how AI governance is technically specified.

Product overview

SaferAI is a French non-profit that creates AI risk measurement tools rather than a commercial product company. The organization operates three integrated offerings: the AI Company Risk Ratings Platform, which independently evaluates and rates AI companies on their risk management practices; a Quantitative Risk Modeling Framework that translates AI capabilities into measurable harm estimates; and advisory services helping companies develop Frontier Safety Frameworks compliant with emerging regulations. The organization also leads AI risk management standards development internationally.

Differentiator

Problem solved

Functional benefit

Products and services

• AI Company Risk Ratings Platform Public platform at ratings.safer-ai.org that independently evaluates and rates how well leading AI companies manage risks across four dimensions: how they identify risks, assess their severity, implement safeguards, and govern their systems. The ratings create market accountability through media coverage, market pressure, and investor decisions, and are referenced by institutional investors and policymakers for sourcing decisions.
• Frontier Safety Framework Advisory Services Direct advisory engagements helping AI companies develop and implement Frontier Safety Frameworks compliant with emerging regulations such as California's SB53 and the EU Code of Practice for general-purpose AI models. Service helps companies develop compliant safety documentation and integrate AI risk assessment into decision-making.

Quantifiable outcome

• Nine risk models developed for AI-enabled cyberattacks
• +1 more outcomes

Named customers3 records

Segments3 records

Ideal customer profiles3 records

Technology focussed Yes

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

AI capability3 records

Feature3 records

Partnerships

Seven partnerships are on record, tiered core, strategic and minor.

• NIST AI ConsortiumcoreStrategic or Co-development PartnerSaferAI participates in the NIST AI Consortium as a partner, creating direct pathways from research to policy action. The partnership enables integration of SaferAI's quantitative risk modeling approaches into national AI evaluation frameworks.
• OECDcoreStrategic or Co-development PartnerSaferAI was the only AI safety organization invited to the OECD taskforce working on the reporting framework of the Hiroshima AI Process. James Gealy contributes to OECD G7 Hiroshima AI Process reporting framework. Partnership creates direct pathways from funding to policy action.
• Google DeepMindcoreStrategic or Co-development PartnerCollaborated with Google DeepMind researchers on goal-directedness evaluations. SaferAI conducted an independent external review of DeepMind's public safety case for Gemini 2.5 Pro using Assurance 2.0 methodology, identifying gaps and providing recommendations.
• G42strategicStrategic or Co-development PartnerSaferAI contributed to G42's Frontier Safety Framework, providing expertise in AI risk management standards and helping shape the company's approach to responsible AI development.
• AI Safety InstitutescoreStrategic or Co-development PartnerSaferAI advises national AI Safety Institutes across the globe on implementing quantitative risk modeling. National AI Safety Institutes are actively integrating SaferAI's approaches into their evaluation frameworks.
• UK DSIT (Department for Science, Innovation and Technology)strategicStrategic or Co-development PartnerBruno Galizzi spent over three years at UK's DSIT contributing to landmark AI policy initiatives before joining SaferAI as Senior Policy Associate.
• EURAIminorStrategic or Co-development PartnerChloé Touzet leads engagement with civil society organisations, including EURAI, as part of SaferAI's external stakeholder engagement strategy.

Scale indicators2 records

Recent moves6 records

Expansion highlights6 records

Direct peers

• GovAI: Research organization working on AI governance questions for policymakers. Shares SaferAI's mission of producing rigorous, evidence-based research that informs AI regulation and frontier AI company policy.
• Monitaur: AI governance, risk, and assurance platform offering audit, monitoring, and policy tooling for ML systems. Comparable to SaferAI's governance and safety-case review offerings, delivered as enterprise software.
• Apollo Research: AI safety research organization focused on scheming evaluation and frontier model risk assessment. Closely aligned with SaferAI's quantitative risk modeling mission and produces independent evaluations of frontier model behavior.
• Holistic AI: Venture-backed AI governance, risk, and compliance platform serving enterprises with AI assessments and policy tooling. Directly comparable to SaferAI's ratings and advisory offerings, though delivered as a commercial SaaS platform rather than a non-profit service.
• Modulos: AI governance and compliance platform that supports organizations implementing responsible AI practices. Operates in the same AI GRC category as SaferAI's advisory and ratings work, serving enterprise and regulatory audiences.
• METR (Model Evaluation and Threat Research): Non-profit research organization developing evaluations for frontier AI models, particularly autonomous task capability assessments. SaferAI researcher Ben Snodin was previously contracted with METR, and both organizations produce technical AI risk evaluations cited by labs and policymakers.
• Credo AI: AI governance software platform that helps organizations measure, manage, and govern AI risk across use cases. Closest commercial analog to SaferAI's ratings and governance framework work, with overlapping enterprise buyer personas.
• Trustible: AI governance and compliance platform that helps organizations operationalize responsible AI policies and evidence. Operates in the same enterprise AI governance category SaferAI serves through advisory and ratings work.

Broad incumbents

• Deloitte (AI Institute / Risk & Regulatory): Global consulting firm with established AI governance, risk, and regulatory practices advising enterprises and governments. Overlaps with SaferAI's advisory and policy work but as part of a much broader services portfolio.

Emerging players

• Center for AI Safety (CAIS): Research and advocacy organization focused on AI existential risk, including the famous 'Statement on AI Risk' signed by hundreds of researchers. Adjacent to SaferAI in mission but more research- and advocacy-oriented versus SaferAI's quantitative, standards-driven approach.

Market position

Strengths4 records

Weaknesses4 records

Competitive moat4 records

Key risks5 records

Key highlights6 records

Customer concentration

• Xx.com/SaferAI_org
• LinkedInlinkedin.com/company/saferai

Revenue estimate

Valuation estimate

Number of profiles

Profiles21 records

Funding overview

Funding rounds

Investors

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about SaferAI