One of ELOQUENCE’s central goals is to ensure that advances in conversational AI benefit all European languages, not only those with abundant data and resources.
In this interview, Alessio Brutti from Fondazione Bruno Kessler shares FBK’s perspective on the first two years of ELOQUENCE, reflecting on progress in multilingual speech inputs for open-source LLMs, collaboration within the consortium, and the importance of shared platforms such as the Interactive Playground. His insights highlight how fundamental research, open tools, and ethical safeguards come together to support inclusive and responsible AI development.
Q: Looking back at the first two years of ELOQUENCE, which achievement from your team are you most proud of?
I’m proud of the work we’ve done on multilingual speech inputs for open-source LLMs and the impact this work has had on improving performance for low-resource languages, which is something we care a lot about. Of course, there’s still plenty to do and lots of ideas we want to explore, but so far the results have been very encouraging, both in terms of the research we’ve published and the artifacts we’ve created in an area that’s really core to our group.
Q: What has been the most rewarding or inspiring part of collaborating with the ELOQUENCE consortium?
What I’ve found most rewarding is how well-balanced the consortium is. Many partners work in similar research and technology areas, but almost always in a complementary way, with very little overlap. That’s created a really intense and fruitful collaboration, sharing ideas, results, baselines, and knowledge. It feels like everyone brings something unique to the table, which makes the whole experience very inspiring.
Q: Which result, insight, or technological advancement do you feel has had the strongest impact on the project so far?
I think the release of the Interactive Playground at month 18 has had a really strong impact. It gave partners a common platform to integrate their results and technologies, and it will be very useful for our pilot-related demonstrators. Another key achievement is the ethical and legal assessment developed under WP6. That’s been crucial because it ensures our pilots and technologies comply with regulations and align with European values. And more recently, the release of S-Dialog has already started helping pilots generate dialogues and data for their use cases, which is a big step forward
Q: Which moment, milestone or breakthrough from the past two years stands out as especially meaningful for your organisation?
One milestone that really stands out is the release of the multilingual projector for open-source LLMs. It’s a solid, though still preliminary, result toward our goal of providing speech language models for all European languages, a core activity for the Speech Technology Lab at FBK. There’s still a lot of work ahead, but this step is an important foundation for achieving that vision.
Q: What aspect of the project’s final year are you most excited about and why?
The final year is a critical phase for the project, as it focuses on consolidating the research outcomes and translating them into tangible results within our pilots. It’s also the moment to demonstrate the overall impact, both scientifically and technologically, ensuring that the work delivered aligns with the project objectives and creates measurable value.
ELOQUENCE’s strength lies in its ability to combine cutting-edge speech and language research with a highly complementary consortium and a strong focus on real-world impact.
As ELOQUENCE moves into its last year, the focus on consolidating results and demonstrating value through pilots underscores the project’s ambition: to translate scientific advances in multilingual and low-resource speech technologies into tangible, trustworthy solutions that align with European values and serve diverse linguistic communities.
