ELOQUENCE Project Mid-Term Update: Advancing Europe’s Trustworthy Multilingual Conversational AI

ELOQUENCE, a Horizon Europe project, enhances conversational AI for safety-critical applications such as emergency response and healthcare. Supported by the European Commission, it boosts AI reliability across multiple languages and aligns with EU ethical standards to improve human-machine interactions.

[Barcelona, Spain] – Horizon Europe-funded ELOQUENCE project, dedicated to developing trustworthy, multilingual, and bias-free conversational artificial intelligence (AI) technologies, reached a key milestone with its plenary meeting hosted by Telefónica Innovación Digital in Barcelona. Gathering the consortium’s 16 partner organisations, the meeting marked the midpoint of the project’s ambitious three-year journey (2024-2026).

Since its launch in January 2024, ELOQUENCE has made significant strides towards creating adaptive, transparent, and ethically robust AI systems designed to operate effectively across diverse European languages and cultures. The plenary meeting provided an essential forum for partners to showcase their progress, address upcoming deliverables, and strategise on pilot integration and technological advancements for the second half of the project.

Key updates shared during the plenary include advancements in multilingual AI models and their applications in both high-risk safety-critical environments. Project partners reported progress on the use cases:

• Privacy-Preserving Home AI: Developing decentralised learning methods to ensure users’ data privacy and security.
• Social Context-Aware Language Models: Implementing measures to identify, manage, and reduce biases in conversational AI systems, promoting fairness and inclusivity.
• Self-Grounded Virtual Agents: Enhancing AI agents to better understand enterprises’ context and users’ intentions, combining enterprise specifics with pretrained knowledge.
• AI-Assisted Healthcare Call Centre Support: Integrating AI supervision for medical service dialogues, improving the efficiency and quality of customer interactions.

ELOQUENCE Coordinator Jordi Luque emphasised: “Ensuring ethical practices and sustainability are embedded in our AI development process, actively engaging with and understanding societal implications to maintain broad acceptance of our work. In our research lines, we have achieved significant milestones in enhancing AI reliability through innovative methodologies and promoting multimodal and multilingual AI that can comprehend and process diverse cultural nuances effectively. In addition, building strong relationships with other research communities and stakeholders is essential for our project’s success. We have actively sought strategic collaborations and valuable synergies with sister projects like TrustLLM and MEETWEEN, and contacted several relevant initiatives.”

In the second half of the project, ELOQUENCE aims to intensify engagement with industry stakeholders and strengthen Europe’s position in sustainable and inclusive AI. Regular capacity-building webinars, podcasts, and interactive workshops have been announced to foster deeper industry and societal engagement.

The plenary session concluded with an overview of future activities, focusing on technical integration and preparation for the project’s mid-term review scheduled for September 2025. ELOQUENCE remains committed to shaping AI technologies that align with European values, ethics, and standards, ensuring innovation in conversational AI contributes positively to society and industry alike.

For further details on the project and its developments, visit eloquenceai.eu.