SPACE 2

Can AI models engage in realistic conversations with people and perform complex tasks (behaving as a specific persona, such as a client in a psychologist's session)? And how can we combine generative AI tools to create a credible embodied conversational agent?

A historic peak in momentum for Natural Language Learning (NLL) and Conversational AI has emerged recently, spearheaded by the popular GPT-4 technology, capable of overcoming the limits of the initial SPACE project (PWO 2018-2020). In collaboration with other revolutionary AI tools simplifying the creation of credible and realistic avatars, this could revolutionize conversational skills training and digital role-playing formats in education and lifelong learning. This will empower the researchers from the departments of DAE (Digital Arts and Entertainment) and TP (Toegepaste Psychologie) not only to achieve the ambitious goals set by the SPACE project but also to expand upon them.

In SPACE 2.0, we anticipate providing a more functional and credible demo swiftly, thanks to the latest NLL and AI tools. Through its new pipeline, this demo will be easily scalable to multiple fields and domains.