Tensor AI Solutions
Making AI more transparent and sustainable
Tensor AI Solutions is a spin-off of the University of Ulm. The startup specializes in explainable and reliable AI applications. The systems are individually tailored to customers and already fulfill future legal requirements of the EU according to the so-called AI Act.
To make AI safe and trustworthy, the European Union has launched a legal regulation to govern the development and use of AI: the EU AI Act. This regulation, which is expected to take effect next year, bans applications that pose unacceptable risks. Applications that are associated with high risks are subject to strict requirements in terms of transparency and traceability. This is exactly where the startup Tensor AI Solutions comes in: The spin-off company of the University of Ulm has developed a transparent and explainable AI technology that already meets the required transparency obligations.
Application areas and fields for the use of AI that fit into this »high-risk« category range from autonomous driving and automated lending to AI-based diagnostic or therapeutic procedures in medicine. »Once the AI regulation is in place, comprehensive quality assurance and risk management systems will be needed for the development and use of AI in the high-risk category. Specifically, this means: Decision-making processes and data quality must be continuously documented and made transparent,« explains Florian Schinnerling, CFO of Tensor AI Solutions.
As a rule, conventional AI methods such as neural networks cannot guarantee transparency and explainability. »With our AI technology, it can be traced at any time on which basis the system makes decisions and in which way individual decisions were influenced by various parameters,« says Dr. Timo Felser, CEO of Tensor AI Solutions. In this way, undesirable tendencies or »biases« of automated systems can be transparently clarified and prevented. AI-supported systems are already being used in personnel selection processes, for example, and it is not always possible to trace the criteria according to which selection decisions were made here.
The focus of Tensor AI Solutions is currently still on retail, process industry and production, although the company's AI technology can easily be applied in other industries. Another user benefit of the startup's AI applications is the efficient use of resources. Transparent AI analysis reveals which data is actually relevant to a process, saving up to 80 percent in resources. Less computing and storage power is needed for the same computing task, which not only speeds up processes but also saves electricity and energy. The startup's service package also includes individual system integration of customized AI, in which performance is specifically adapted to the hardware and optimized.
