Building trust on understanding... Follow us on social media by searching for THEMIS 5.0
Different people. Different values. Different trust.
Trust means different things to different people. A doctor may prioritise safety and explainability. A journalist may value transparency and accountability. A port manager may depend on reliability and performance. There is no single definition of trustworthy AI because trust is shaped by people's roles, responsibilities, values and the decisions they need to make.
In an era of rapid AI adoption, THEMIS 5.0 helps organisations, developers, decision-makers and professionals understand, assess and improve AI systems based on what matters most to the people who use them, rely on them, or are affected by them.
Combining human-centred principles with risk-based assessment methods, practical tools and structured processes, THEMIS helps organisations evaluate how AI systems work, what data they use, how decisions are made, and where improvements are needed. Rather than asking whether an AI system is simply "trustworthy", THEMIS helps determine whether it is trustworthy for a specific context and the people within it.
The result is more informed, transparent and accountable AI decision-making. Organisations can move beyond experimentation to deploy AI with greater confidence, addressing real operational and societal challenges while building systems that people understand and trust.
OUR MISSION
Trustworthy AI Starts With People
16
Partners
8
Countries
36
Months
I Use AI
Can you trust your AI for the decisions you need to make?
I Build AI
How do you build AI that people actually
trust?
I Govern AI
How do we deploy AI responsibly and
confidently
Let Us Help
Fairness in AI includes concerns for equality and equity by addressing issues such as bias and discrimination. THEMIS acknowledges perceptions of fairness differ among AI users and may shift depending on usage of the AI system.
AI-BASED ASSESSMENT OF FAIRNESS
Assessment of technical accuracy and robustness in the ‘AI-based Assessment of Technical Accuracy & Robustness’ module will be implemented on the principles of an AI-driven anomaly detection system.
AI-BASED ASSESSMENT OF TECH ACCURACY & ROBUSTNESS
“AI based assessment of Impact of Decision” will be a qualitative model of the socio-technical environment containing all the possible actions that the AI system can recommend, and all the KPIs that might be affected by any of the possible actions.
AI-BASED ASSESSMENT OF DECISION IMPACT
The THEMIS "AI-based Anonymous Assessment of Human" module will incorporate the human perspective to the evaluation of trustworthiness by assessing both human behavioural and moral value aspects.
