The need for a Fair, Inclusive & Transparent AI certificate
Safeguards against risks
The Fair, Inclusive & Transparent AI (Fit AI) Certificate of Conformity will ensure AI systems meet established ethical and legal standards for fairness, inclusivity, and transparency specifically, protecting developers, users, and society.
Legal compliance
With growing AI regulations, such as the EU's AI Act, our certification will help systems align with laws on fairness, transparency, and data protection specifically, simplifying navigation of complex legal frameworks.
Risk mitigation
Through rigorous testing, the Fair, Inclusive & Transparent AI (Fit AI) certification will reduce bias and discrimination risks and promote fairer, inclusive and transparent AI deployment.
Trust & credibility
The Fair, Inclusive & Transparent AI (Fit AI) certification will build public trust by assuring stakeholders that systems operate ethically, fairly and transparently, fostering adoption.
Market access
In regulated markets like the EU, certification is often mandatory. The Fair, Inclusive & Transparent AI (Fit AI) Fit AI will aim to be a key prerequisite for entry.
Ethical & social responsibility
The Fair, Inclusive & Transparent AI (Fit AI) certification will embed the ethical principle of fairness, aligning AI systems with societal values and minimising harm.
Liability reduction
Certification will demonstrate due diligence, reducing legal risks for developers / producers and companies if issues arise.
Fostering innovation
By providing clear guidelines and benchmarks, the Fair, Inclusive & Transparent AI (Fit AI) certification will encourage innovation while fostering fairness, inclusivity and transparency.
Existing AI certification
Fragmented landscape
ISO / IEC AI Standards
Fragmented landscape
As of now, there is no universal, globally recognised Certificate of Conformity specifically for AI systems. However, there are several emerging frameworks, standards, and certification schemes that aim to establish conformity assessments for AI. These efforts are often driven by governments, international organisations, and industry coalitions. They are challenged with trying to address several critical concerns altogether, such as safety, accountability, and robustness, as well as fairness, inclusivity and transparency. We include some notable examples here.
EU AI Act
ISO / IEC AI Standards
Fragmented landscape
The European Union's AI Act has introduced a risk-based classification system for AI systems:
High-risk AI systems require conformity assessments to ensure compliance with standards related to safety, fairness, transparency, and robustness. AI developers / producers and users / deployers in the EU may need to obtain certifications to prove conformity.
ISO / IEC AI Standards
ISO / IEC AI Standards
IEEE certifiable programmes
The International Organization for Standardisation (ISO) and the International Electrotechnical Commission (IEC) have developed standards for AI, including:
ISO/IEC 42001: A management system standard for AI systems. Other standards address specific aspects of AI, such as trustworthiness (e.g., ISO/IEC TR 24028), bias mitigation, and explainability.
Certification based on these standards is already possible for some implementations through third-party auditors.
IEEE certifiable programmes
IEEE certifiable programmes
IEEE certifiable programmes
The IEEE Standards Association has developed guidelines for ethical and human-centered AI under its Ethics in AI and Autonomous Systems series, such as the IEEE 7000 Series (e.g., bias, transparency, data privacy). Certification programs for compliance with these standards are emerging but are not yet widespread.
CE marking for AI in the EU
IEEE certifiable programmes
CE marking for AI in the EU
Certain AI systems that fall under existing EU directives (e.g., machinery, medical devices) may already require CE marking. This process involves demonstrating conformity to essential requirements, including AI-related aspects.
The EU AI Act expands these requirements to more AI systems.
NIST Frameworks in the US
IEEE certifiable programmes
CE marking for AI in the EU
The National Institute of Standards and Technology (NIST) in the U.S. published its AI Risk Management Framework in 2023. While it is not yet a certification, the framework is widely regarded as a foundation for future conformity standards in the U.S.
Industry-specific certifications
Private sector & non-governmental efforts
Private sector & non-governmental efforts
Medical AI: For AI systems used in healthcare (e.g., diagnostic tools), certifications often align with existing regulatory frameworks like the FDA in the U.S. or the MDR in the EU.
Autonomous Vehicles: Standards like ISO 26262 for functional safety in automotive systems include AI-specific elements, requiring certification.
Private sector & non-governmental efforts
Private sector & non-governmental efforts
Private sector & non-governmental efforts
Organisations such as AI Global, CertifAI, and Algorithmic Accountability Labs are piloting AI certification schemes. Companies like IBM and Microsoft offer self-assessment tools and certification-like programs for trustworthy AI.
Fit AI: shaping the future of AI
Private sector & non-governmental efforts
Fit AI: shaping the future of AI
The existing AI certification landscape is fragmented due to varying regional regulations. However, initiatives like the EU AI Act, ISO/IEC standards, and NIST frameworks have been driving progress. As regulations evolve, standardised AI certifications will become more common.
The Fair, Inclusive & Transparent AI (Fit AI) certification aims to ensure that the risks of bias and discrimination are effectively addressed —without overshadowing critical concerns like safety, accountability, and robustness - and set the global gold standard for AI integrity, ethics and trust.