Smart Data & Knowledge Services

We research human-centered AI systems that are explainable, robust and ready for application.

GeT-AI stands for the development of generative and transparent AI technologies that respond to real needs from society, industry and public administration. The focus is on human-centered solutions that are powerful, explainable and responsible.

Our research combines methodological excellence with a clear application focus - from digital forensics, and Industry 4.0 to innovative solutions for sustainability and ease-of use AI for creatives. The goal is AI that is trustworthy, scalable and socially relevant.
 

Vision and mission

Our vision

We design AI technologies that put people at the center, contribute to societal value and convince through transparency and trustworthiness.

Our mission

Our work systematically integrates transparency, robustness and performance into the entire development process of AI systems. We are guided by clear principles: resource-saving model architecture ("small AI first"), data-minimizing processes and collaborative human-AI interaction. Our solutions are developed in close cooperation with partners from science, industry and the public sector.

Research focus areas

Generative AI and foundation models

• Development of generative, application-specific AI models
• World foundation models for industrial applications of physical AI
• Multi-agent systems for cooperative problem solving

Explainable and transparent AI

• Interpretable deep learning for risk-critical contexts
• Continuous (lifelong) learning based on efficient model architectures
• Tools for human-AI collaboration in decision-making

Digital forensics and security

• Forensic analysis and reconstruction of digital evidence
• Audio forensics and detection of manipulated content
• Classification and detection of malicious digital content

Sustainable AI

• AI-supported analysis from earth observation data
• Applications for agriculture, urban development and disaster prevention

Transfer labs and piloting

KIForPol: Research transfer for the security authorities

With the KIForPol (Artificial Intelligence for Police and Security Authorities) transfer lab, we are building a bridge between basic research and operational application in collaboration with the Federal Criminal Police Agency (BKA) and the Rhineland-Palatinate State Criminal Police Agency (LKA). The aim is to test and further develop new AI methods - for example, for analysing digital traces, detecting manipulated content or supporting police investigations.

In particular, the focus is on AI that can be explained in court - as a contribution to a technologically modern security architecture anchored in the rule of law.
The Transfer Lab is embedded in the federal government's digital strategy and is an example of the co-creative design of trustworthy AI in the public sector.

More about the cooperation with the BKA and LKA

Fields of application

The research results of GeT-AI are used in a wide range of practical fields, including

• Digital forensics and decision support for law enforcement agencies
• Industrial applications and production processes
• Administrative technologies and modernization of public authorities
• Creative media and design processes
• Environmental monitoring and spatial planning