How does Artificial Intelligence influence our life - today and in the future?
Elsa A. Kirchner: AI already accompanies us in everyday life and simplifies our lives. Examples include navigation systems in cars, smartphones or devices in the so-called smart home, such as Amazon Echo. They understand our language and thus enable simple voice input. Some of the underlying AI systems are becoming increasingly familiar with our preferences and interests, particularly through our use of the Internet - usually with the aim of placing individually tailored advertisements. In the future, the health data we collect with various apps on our smartphones may help to support medical diagnoses and improve prevention. Further possible applications result from the recognition of movements and learning of typical movement patterns: The identification of persons, which is already done today via AI-supported systems using face recognition, will thus become even more precise - and can increase security at airports, for example.
What fascinates you about the development of AI?
Elsa A. Kirchner: In my research I focused on the future use of Artificial Intelligence for human-machine interaction, for example in care and rehabilitation. There, I am particularly fascinated by the possibility of learning individual models about humans - for example, about the strengths and weaknesses of the human movement system after an illness or injury. If one learns these models with appropriate data, robotic assistance systems can be adapted online to the individual needs of the user. The assistance systems thus learn with the help of AI which preferences and needs a person has in general or in specific situations - and can provide them with precisely tailored support. An example: An AI-based exoskeleton supports rehab patients only as far as it is actually necessary depending on their daily condition or in a specific situation. It thus creates ideal conditions for rehabilitation. In the future, AI systems will not only learn on the basis of one modality (e.g. image data), but will draw their information from various sources - be it image data, motion data, biographical data or biosignals. This will enable them to make even more precise or individualized assumptions. To develop and validate new methods or approaches for this purpose is a great challenge and fascinating research question.
What qualifications should female graduates have for AI research?
Elsa A. Kirchner: Graduates should absolutely have the desire to look beyond the domain boundaries of computer science. In AI research, it is not enough to be able to program or use frameworks and tools or "only" to develop a better algorithm for a special method of machine learning. It is rather a matter of developing new approaches with comprehensive methods that allow the next big step in AI research.