Safeguarding AI
IABG's safeAI team develops solutions for the evaluation of AI safety and robustness, currently focusing on computer vision applications.
We combine the vast experience of IABG in testing, analyses, and certification processes as well as the latest AI-focused activities in standardization committees like DIN and ISO/IEC with current academic research.
Our software provides a safety framework for autonomous systems.
Aspects of AI Safety
Compliance
In the development of the safeAI safety framework, current AI regulations are taken into account.
Robustness
The safeAI toolbox helps your AI to fulfill state-of-the-art measures on AI robustness.
Data
Data is collected, labeled, and augmented automatically from within our simulation.
Runtime Assurance
The runtime assurance concept is utilized by the safeAI drone simulator.
Deployment
The deployment of machine learning workflows increases efficiency and reproducibility.
Monitoring
The safety dashboard provides meaningful insights into the performance of your algorithm.
Our Solutions
Simulation Environment
Based on Unreal Engine and AirSim, our photorealistic simulation provides support for any kind of sensor possible - to create a world as detailed as reality itself.
Safety Dashboard
Adjust simulation parameters by hand or let our software suggest how to improve your algorithm.
Uncertainty Quantification
Estimate and interpret uncertainty of each prediction. Detect untrustworthy results and ensure safe decision-making.
Applications
Urban Air Mobility
Our software enables a safe flight of autonomous drones and air taxis.
Autonomous Driving
We audit and assess AI models to mitigate vulnerabilities of autonomous vehicles.
Healthcare
safeAI provides explainable and trustworthy AI decisions in medical imaging.