EU Issues EUR 20M RFP for Large Language Models that ‘Respect European Values’

Large Language Model European Values 2023

On December 9, 2022, the European Commission announced a EUR 20m tender for Natural Language Understanding and Interaction in Advanced Language Technologies under the Horizon Europe funding program for research and innovation. The call for proposals will be open until March 29, 2023, followed by an evaluation period expected from April to May.

Proposals “should set out a credible pathway to contributing to a human-centered and ethical development of digital and industrial technologies.” Data science, computer science, computational linguistics, machine learning, and natural language processing are just a few areas where proposals should show their expertise. Special attention should be given to identifying and controlling gender bias — or other biases — in language models. The language models should be “respectful of European values (i.e., privacy, non-discrimination, robustness in legal, ethical and technical terms, reliability and trustworthiness, interpretability and explainability, security and safety).”

Expected Outcomes

As stated in the call for proposals, the advancement of artificial intelligence (AI) presents opportunities for humans to “directly use and benefit from smarter systems.” To achieve effective AI-based human-machine interaction and collaboration, it is crucial to comprehend the real meaning of natural languages, recognize gestures and activities, understand intentions, establish and sustain shared mental models, and design multi-step interactions.

The AI solutions that are envisaged should focus on improving context-aware human-machine interaction and supporting and enhancing human-to-human communication across languages. Therefore, the winning projects are expected to contribute to:

  • developing advanced language technologies for producing context-aware language models,  which can “achieve natural behavior and reasoning in all intended settings”
  • creating effective multilingual and unbiased language models, which are “capable of learning from smaller language corpora” and “respectful of European values”
  • building AI systems based on innovative multilingual pre-trained language models
  • encouraging a higher adoption of innovative language technology solutions by European companies to provide extensive language coverage of AI-enabled applications and services in Europe.
SlatorCon Silicon Valley 2024 | $ 1,340

SlatorCon Silicon Valley 2024 | $ 1,340

A rich online conference which brings together our research and network of industry leaders.

Buy Tickets

Register Now

More specifically, improving context-aware human-machine interaction involves increasing the understanding and exploitation of the interaction context and content in multimodal settings to increase the responsiveness of interactive AI solutions — such as smart assistants, conversational and dialogue systems, and content generation models.

On the other hand, cross-lingual communication can be achieved through automatic translation or interpretation (including automatic subtitling) in real time with a greater understanding of the communication context and meaning involved in it.