Peer
The HyPEr ExpeRt Collaborative AI Assistant
The project at a glance
The PEER project will achieve significant breakthroughs in how to systematically design, realize, and evaluate human-centric AI for sequential decision-making settings.
The project will deliver AI solutions for such settings while putting the user at the centre of the entire design, development, deployment, and evaluation pipeline, leading to truly mixed human-AI initiatives.
The aim is to enable a more helpful bidirectional conversation between humans and AI, more fluid mutual learning and reasoning, and more productive collaborative work, which in the end allows for increased user trust and acceptance.
To achieve this, PEER will create a cross-discipline symbiosis between social sciences and artificial intelligence.
The following sub-objectives have been identified to achieve the overall ambition and objective of PEER:
- defining a human-AI interaction design method
- enhance responsiveness and adaptability of AI systems for sequential decision problems
- measure progress towards truly mixed and trustworthy AI
- integrate and demonstrate the PEER collaborative AI assistant in real-world domains
- enhance the adoption of developed techniques by connecting to the AI on demand platform, PPP AI data & Robotics
CATIE contribution
CATIE will lead the WP4 to provide the Artificial Intelligence Acceptance (AIA) index to allow the benchmarking of trustworthy AI application.
Main objectives:
- Defining a set of measurement scales
- Designing the tools and assessment processes associated
- Providing a common framework for the evaluation and assessment of the AI system
- Manage and organize the evaluation and assessment process in all the use-case/pilots
CATIE will also provide an IHM for the interpretation of the AIA index.
Expertise used
This project has received funding from the European Union’s Horizon Europe
research and innovation programme under grant agreement No 101120406