Establishing a new era in human-AI collaboration
As artificial intelligence (AI) becomes increasingly ingrained in our everyday lives, it is critical that the transformative power of this technology is fully directed at benefiting citizens. Finding ways of cementing this was the central aim of the EU-funded HumanE-AI-Net project. “The project’s vision was to look at AI not as a means of doing what humans already do but better, but rather as a way of collaborating to enhance human capabilities,” explains Paul Lukowicz, the network coordinator from the German Research Center for Artificial Intelligence. “The idea is that AI enables you to do things you want to do, but in a stronger and more effective way, a bit like having a ‘cognitive exoskeleton’.”
AI, human-computer interactions and social science
To achieve its aims, the project brought together people with core AI expertise, as well as experts in the fields of human-computer interaction, social sciences and law. The goal was to solve technical problems related to making human-AI interaction beneficial to individuals and society. “We need to understand how AI affects the way we live together,” says Lukowicz. “This needs to go into the design process, rather than just being something that is imposed afterwards.” The project was organised into a number of interdisciplinary micro-projects. These brought together a diverse range of researchers and sought to tackle specific issues related to human-AI interaction. This resulted in a number of deliverables including an impressive number of publications and reports.
Practical AI tools to support society
A number of micro-projects were also successful in delivering practical tools. One initiative for example developed an augmented reality system to support pancreatic surgeons in removing tumours. “The AI system takes images of the tumour and surrounding tissue and puts them into augmented reality glasses that surgeons wear while operating,” explains Lukowicz. “The operation is effectively a collaboration between AI and the surgeon.” Another micro-project looked to leverage AI for more accurate and nuanced air quality analysis. The aim was to take account of pollutants that can be highly variable, such as road traffic, by harnessing the power of AI and big data rather than through using more expensive roadside sensors. Other initiatives looked at developing new tools for combating social media bias and disinformation. “Chatbot technology to help smooth interactions between migrants and bureaucracies was another very practical application to emerge from the project,” notes Lukowicz. “Filling in documents and applications takes time, and misunderstandings and mistakes can be made. Researchers looked at ways to help make this process smoother.”
Research community for human-AI collaboration
Another key element of the project was the establishment of a sustainable research community in the field of human-centric AI. An international conference series was developed and launched in collaboration with the Dutch Hybrid Intelligence Centre to create a space for discussion and collaboration. The annual Hybrid Human-Artificial Intelligence conference was first held in 2022, with the fourth edition scheduled for June 2025 in Pisa, Italy. “Scientific communities are usually defined by the venues where they publish and meet,” adds Lukowicz. “This conference series, which will continue after this project, will help to create a new community around the notion of human-AI collaboration.” The project team is also in the process of publishing a handbook on collaborative AI, which the team hopes will become a reference for practitioners in the field.
Keywords
HumanE-AI-Net, AI, societal, surgeons, big data, air quality
