The Project

The chemical industry is characterized by complex logistical processes, whose efficiency and effectiveness are crucial for economic success. These processes involve both the logistics of materials and products, as well as the logistics of employees. Complete automation of these processes is not always feasible or desirable, as there are too many exceptions that companies must respond to flexibly. However, the input parameters necessary to understand these processes are so extensive that employees cannot fully utilize the information they contain.

Artificial Intelligence (AI), on the other hand, is highly effective at recognizing patterns in data streams. However, AI typically cannot assess all the potential real-world impacts of its results. Therefore, the hKI-Chemie (human-centric AI) project explores how AI can empower people to make better decisions in complex situations. To achieve this, AI solutions are being developed that are not black boxes but are self-explanatory. This approach enables the chemical industry to achieve multiple goals simultaneously:

• Optimization of processes through AI-based evaluation of process parameters
• Support for employees in early detection of process issues and in selecting appropriate solutions
• Availability of machine-learned insights across shifts and personnel changes

The goal is not for AI to replace humans. Instead, the project investigates how AI can assist people in making well-founded and optimized decisions. This aligns with the vision of human-centered AI, as outlined in the German government’s AI strategy and the final report of the Enquete Commission on “Künstliche Intelligenz – Gesellschaftliche Verantwortung und wirtschaftliche, soziale und ökologische Potenziale”.

hKI-Chemie collaborates closely with the social partners of the chemical industry, as AI methods will have a fundamental impact on modern workplaces. In addition to employer associations and trade unions, the works councils of the participating chemical companies are actively involved in the project. The project aims to develop reusable hKI modules, development methods, and action recommendations to facilitate the adoption of human-centered AI across various companies and industries.

Links:

• Website

---

Publications (3)

2024 (1)

• Zdankin, Peter and Weis, Torben
  Demo: Smartbright: Learning a Physics-Based Model of Light-Scattering With Symbolic Regression
  In: Proceedings of the 2024 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2024 ACM International Symposium on Wearable Computers

2023 (2)

• Becks, Eileen and Josten, Malte and Matkovic, Viktor and Weis, Torben
  Revising Poor Man's Eye Tracker For Crowd-Sourced Studies
  In: 21st IEEE International Conference on Pervasive Computing and Communications (PerCom 2023)
  DOI: 10.1109/PerComWorkshops56833.2023.10150386
• Becks, Eileen and Josten, Malte and Matkovic, Viktor and Weis, Torben
  Artifact: Revising Poor Man's Eye Tracker For Crowd-Sourced Studies
  In: 21st IEEE International Conference on Pervasive Computing and Communications (PerCom 2023)
  DOI: 10.1109/PerComWorkshops56833.2023.10150307