AI systems are increasingly used in areas where their decisions can have a direct impact on people and organisations. This makes it important not only to assess their technical performance, but also to understand whether they meet the needs and expectations of the people who use them. Through THEMIS 5.0 and FAITH, ATC has been working on this challenge from different perspectives, focusing in particular on the role of the user in the assessment of AI trustworthiness.

THEMIS 5.0 Human-centered Trustworthiness Optimisation in Hybrid Decision Support, is a 3-year HE RIA project that started in 2023 and that draws researchers and practitioners from diverse disciplines to implement an AI-driven, human-centred Trustworthiness-Optimisation ecosystem following an AI-driven risk assessment approach. In this project, ATC developed the Persona Analyser, a conversational agent that drives a conversation with the end-users to identify their trustworthiness preferences in terms of prioritising the AI system’s Fairness, Accuracy and Robustness. The Persona Analyser also acts as an intermediary, translating model-level technical concepts into non-technical, real-world examples, enabling users to express granular trustworthiness requirements in ways that can be mapped to AI model characteristics. 

Figure1. ATC’s results from the THEMIS 5.0 and THEMIS projects 

Building upon THEMIS 5.0, FAITH Fostering Artificial Intelligence Trust for Humans towards the optimization of trustworthiness through large-scale pilots in critical domains, is a 4-year HE IA project that started in 2024 and that to develop and validate a human-centric, trustworthiness optimization ecosystem, which enables measuring, optimizing and counteracting the risks associated with AI adoption and trustworthiness in critical domains by adopting a dynamic risk management approach following EU legislative instruments and ENISA guidelines. In this project, ATC developed TrustGuard, a checklist-based tool for implementing a risk-based methodology, compliant with ISO27005 and ISO42001, that aims to assist a risk assessor to assess the threats and estimate the underpinning risks of all components of an AI system during its entire lifecycle (design, development, deployment). 

Challenges 

Experience from THEMIS 5.0 and FAITH highlighted the importance of addressing the human factors of trustworthy AI in socio-technical systems and strengthening Human-in-the-Loop (HITL) components so that AI behaviour and optimisation priorities reflect user needs across sectors and use cases. Key challenges include: 

• Technical Parameters interpretation: Non-technical users often struggle to understand model-level technical details and trade-offs. As a result, they cannot confidently prioritise technical improvements that best align the AI system with their needs. 

• Use Case context: Translating the use-case context (e.g. constraints, requirements, and business targets) into measurable KPIs connected to model parameters is difficult. Since this process relies heavily on user input, it can be affected by bias and limited visibility of the interdependencies involved.    

• GenAI: Many trustworthiness assessment frameworks and tools still face difficulties in analysing Generative AI systems. This gap needs to be addressed given the rapid uptake of GenAI tools and the distinct risk patterns they introduce (e.g., content reliability, hallucinations, misuse, and provenance). 

Towards a human-centric AI Trustworthiness Assessment Framework 

Building on knowledge and results from THEMIS 5.0 and FAITH, ATC aims to advance its HITL capabilities towards a human-centric AI Trustworthiness Assessment Framework. The proposed direction is an agentic approach that combines multiple AI paradigms (e.g., GenAI, machine learning, and symbolic AI) to integrate and operationalise different aspects of AI trustworthiness.  

In particular, the approach will: 

1. Extract and represent domain knowledge from use-case documentation and datasets, capturing context, constraints, and objectives. 

1. Translate technical trustworthiness parameters into non-technical, real-world examples linked to business objectives, risks, and measurable KPIs, so users can express priorities in actionable terms. 

This enables non-technical users to define use-case-specific technical requirements for improving trustworthiness, and to identify missing trustworthiness controls and improvement actions. By combining these outputs, the framework can also support a preliminary risk assessment, highlighting non-implemented controls that have the highest expected impact on trustworthiness for a given use case. 

Figure2. Agentic Human-in-the-Loop component for a human-centric AI Trustworthiness Assessment Framework

Overall, the proposed framework builds on the work already carried out in THEMIS 5.0 and FAITH and focuses on making AI trustworthiness assessment more useful for non-technical users. The main goal is to better connect user needs and use-case requirements with technical parameters, controls and risks, so that the assessment can lead to more concrete improvement actions.

During ATC’s latest company-wide meeting, held on 9 July 2026 at the company’s offices, Eleni Tsalapati presented Personal Companion for Countering Disinformation, an AI-powered system developed by ATC as part of the EU-funded AI-CODE project.

Personal Companion is designed to support journalists, fact-checkers, and other media professionals in assessing the credibility and factual accuracy of online content. It analyses articles by cross-referencing claims with trusted external resources, detecting potential flaws in reasoning and hate speech, and estimating the impact and spread of claims across social media.

What sets Personal Companion apart from standard foundation models is its trustworthiness. The system has undergone extensive evaluation to ensure high levels of accuracy, robustness, and fairness. It has also been designed to be efficient, interpretable, and to keep humans in the loop, empowering professionals rather than replacing them.

Claim verification through cross-referencing with trusted external resources.

The feedback received during the presentation was particularly encouraging, with media professionals estimating that the tool could reduce the time spent verifying content by more than 30%.

The technologies developed for Personal Companion have already been reused in two additional EU-funded projects, AI-REFRAME and BROD, demonstrating their versatility and impact. At the same time, work carried out within the FAITH project continues to strengthen the system’s trustworthiness.

The development of Personal Companion shows how research outcomes can evolve into reusable technologies that continue to create value across multiple projects and initiatives.

CIRCAT Open Call 1 webinar registration is now open. The webinar is designed for organisations interested in applying to CIRCAT Open Call 1, focused on Penetration Testing Scenarios Development for Europe’s critical infrastructures.

The CIRCAT project (Cybersecurity Infrastructure Resilience, Collaboration, and Advanced Training) launched its Open Call 1 (Penetration Testing Scenarios Development) to support European organisations in strengthening cybersecurity resilience by defining and documenting realistic cyber threats against critical infrastructures.

CIRCAT is funded by the European Union under Digital Europe Programme (DIGITAL). The project aims to enhance cybersecurity resilience for large industrial installations and sectors indicated as critical infrastructure sectors in NIS2 (Directive (EU) 2022/2555) by leveraging AI-driven tools and processes.

SCAF-DOME is a short-term initiative funded under the Horizon Europe Programme and linked to the CyberSecDome framework, which advances cybersecurity through AI-enabled security capabilities and immersive incident investigation and response concepts.

The FARM360TECH project is a transformative project that aligns with the I3-2024-INV2a call, advancing the digital and green transformation of European agriculture.

Disinformation about migrants and refugees doesn’t just spread falsehoods. It spreads fear. And fear fuels division. AI-REFRAME is a bold new project designed to stop that. It uses cutting-edge AI to detect and counter online misinformation and hate speech targeting migrants and refugees.