Trustworthy AI for a Sustainable Future: How THEMIS 5.0 Supports the Sustainable Development Goals

Artificial intelligence (AI) is increasingly helping us solve some of society's greatest challenges. It is supporting doctors to diagnose disease, helping journalists verify information, improving transport and logistics, and enabling organisations to make faster, more informed decisions. Yet AI alone will not help us achieve the United Nations Sustainable Development Goals (SDGs).

The United Nations Sustainable Development Goals (SDGs) are 17 interconnected global goals that provide a shared framework for creating a more sustainable, equitable and prosperous future. Trustworthy AI has an important role to play in helping achieve many of these ambitions.

For AI to deliver meaningful social, economic and environmental benefits, people must be willing to trust it. That trust, as with all types of trust, cannot be assumed. It must be earned. This is the challenge at the heart of our Horizon Europe THEMIS 5.0 project.

Rather than developing another AI application, THEMIS brought together 16 partners from across Europe to create a human-centric Trustworthiness Optimisation Ecosystem that helps organisations assess, understand and improve AI systems before they are deployed. By combining technical innovation with ethics, law, human behaviour and real-world user experience, our project is helping ensure that AI contributes to society in ways that are responsible, transparent and sustainable.

Sustainable AI Starts With People

One of the defining characteristics of THEMIS is that the project did not begin by asking 'How should we assess AI?' Instead, it asked a more fundamental question:

'What do people need before they trust AI?'

From the very beginning, users were placed at the centre of the project. More than 270 participants across eight countries took part in co-creation activities, Living Labs and stakeholder workshops, helping the consortium understand how different people perceive trustworthiness in different contexts. These activities informed citizen co-created design principles, user personas, platform requirements and the overall methodology behind the THEMIS ecosystem.

This was an important shift. Rather than designing technology first and asking users to adapt, THEMIS used co-creation to ensure that technology was shaped by the needs, concerns and expectations of the people who will ultimately use it. A philosophy runs throughout the entire project.

Trust Is Not One-Size-Fits-All

A recurring lesson from the project is that trustworthiness means different things to different people. A clinician evaluating an AI system for medical imaging needs confidence that the system is accurate, explainable and safe before incorporating it into patient care. A journalist using AI to analyse information wants transparency, traceability and accountability before publishing a story. A port operations manager considering an AI system that predicts vessel arrival times needs reliable forecasts that continue to perform under changing weather conditions and operational disruption.

Each person is assessing the same AI system through a different lens. THEMIS recognises this reality as a foundational principle of our work. Our platform combines quantitative trustworthiness assessment with user context, risk assessment and explainable decision support, helping users understand the evidence most relevant to their role while still considering the complete trustworthiness picture. This human-centred approach moves beyond a single trust score and towards informed decision-making.

Co-Creation as a Driver of Sustainable Innovation

Collaboration sits at the heart of sustainable development, and therefore it also sits at the heart of THEMIS. Our consortium combines expertise in artificial intelligence, cybersecurity, law, ethics, human-computer interaction, psychology, communications and risk management. Universities, research institutes, SMEs, technology providers and end users have worked side-by-side throughout the project, ensuring that technical innovation is built and tested against real operational needs.

This multidisciplinary approach is reflected in THEMIS's three pilot environments.

In healthcare, clinicians explore how AI can support diagnosis while ensuring that medical professionals retain responsibility for clinical decisions. In port operations, AI helps predict vessel arrival times to improve berth allocation, cargo handling and workforce planning, while THEMIS provides evidence about whether the AI is sufficiently trustworthy for operational deployment. In media, journalists evaluate AI-supported information while maintaining professional standards of transparency, accountability and editorial responsibility.

Across all three pilots, the same lesson emerges: Technology alone does not create trust, people do.

Trustworthy AI From Principles to Practice

Many trustworthy AI frameworks describe what good AI should look like. THEMIS goes one step further by helping organisations operationalise those principles. One of the project's major innovations is the Trustworthiness Optimisation Process (TOP), a structured methodology that supports organisations throughout the AI lifecycle. Rather than treating trustworthiness as a one-off assessment or compliance exercise, TOP enables continuous evaluation, improvement and optimisation of AI systems as they evolve.

Supporting this methodology is the integrated THEMIS platform, which combines trustworthiness assessment, user profiling, explainable decision support, risk assessment and continuous improvement into a single ecosystem. Together, these components help organisations understand not only how an AI system performs, but whether it is ready for deployment within their own operational context.

Contributing to the Sustainable Development Goals

This human-centred approach means THEMIS contributes to several Sustainable Development Goals (SDG's).

SDG 3 – Good Health and Well-being

THEMIS supports safer adoption of AI in healthcare by helping clinicians understand the capabilities, limitations and trustworthiness of AI-assisted decision support before it is used in practice.

SDG 5 – Gender Equality

The project recognises that trustworthy AI must also be inclusive AI. Through co-creation, participatory technology assessment and diverse stakeholder engagement, THEMIS encourages AI systems that consider different perspectives and helps identify potential bias before deployment. The consortium itself reflects this commitment through its near-equal gender balance, female scientific leadership and expertise in gender dimensions of AI.

SDG 9 – Industry, Innovation and Infrastructure

THEMIS supports innovation by helping organisations deploy AI with confidence. In the port pilot, trustworthiness assessment enables operators to evaluate AI systems before they become part of critical infrastructure operations.

SDG 16 – Peace, Justice and Strong Institutions

Through explainability, accountability and human oversight, THEMIS promotes responsible AI governance, helping organisations move beyond compliance towards genuine transparency and informed decision-making.

SDG 17 – Partnerships for the Goals

Perhaps the project's greatest contribution is its collaborative model. THEMIS demonstrates how researchers, technology developers, policymakers and end users can work together to co-create AI systems that are technically robust, socially acceptable and aligned with European values.

Building AI That People Can Trust

Like trust itself, the Sustainable Development Goals cannot be achieved through technological innovation alone. They require technologies that people understand, question and ultimately choose to adopt. THEMIS shows that trustworthy AI is not simply about improving algorithms. It is about involving people from the very beginning.

It is about recognising that trust differs between sectors, professions and individuals.

It is about combining technical excellence with ethics, transparency, explainability and continuous learning.

Most importantly, it is about ensuring that AI remains a tool to support human decision-making, not replace it. As Europe continues to invest in digital transformation, projects like THEMIS demonstrate that sustainable AI is not defined solely by what technology can do, but by how responsibly it is designed, assessed and used. Because the future of AI is not just about building smarter systems, it is about building the systems that society can actually trust.