Potentials and limits of artificial cognitive systems
Artificial cognitive systems are IT systems. They can sense, interpret and actively modify both their physical and virtual environment in order to generate solution strategies based on artificial intelligence (AI) for a wide range of tasks. Although such systems have already demonstrated their enormous potential, this can only be fully realised in collaboration with humans. The working group deals with different architectures of artificial cognitive systems, their characteristics and performance as well as their possible applications and limitations. Another important focus is the comparison with the function and structure of natural cognitive systems. The topic is attracting broad interest. Accordingly, it requires a particularly interdisciplinary perspective due to the current and future areas of application, the wide range of possible uses and social implications.
Goals
➤ Fundamental research questions are:
- What exactly is a natural and what is an artificial cognitive system?
- What are the functional principles of cognition?
- What exactly does ‘intelligence’ mean in such systems?
- What role does physicality (embodiment) play in cognition and intelligence?
- How similar can or should AI systems be to the architectures of natural cognitive systems?
- What are the prerequisites for artificial creativity or artificial consciousness and what are the premises of attribution?
- What roles do images of human beings play in this?
➤ Illuminating the applications and limits of artificial cognitive systems
- What are the new potential applications of cognitive systems, for example in medicine or law?
- What possibilities do AI systems offer for replacing or supplementing currently still inherently human abilities such as creativity or empathy?
- What role will AI play in research?
- What are the technical and ethical limits?
➤ Exploring interaction with human intelligence
- In what ways will we interact with AI systems?
- What conditions must be met for us to trust AI systems?
- How is dealing with AI learnt, for example at school and at universities?
Project Description
Artificial cognitive systems as intelligent IT systems range from software agents (e.g. ChatGPT from OpenAI) to cyber-physical systems (e.g. Phoenix from Sanctuary AI). They can independently develop solutions for different tasks. To do this, these systems must perceive their real and digital environments multimodally, draw conclusions and learn. They can also change their environment if necessary. AI methods such as machine learning, neural networks and deep learning are used here. Interfaces to the real and digital world are necessary for interaction with the environment. While earlier artificial cognitive systems often had to rely on robotic and cyber-physical systems with complex sensors and actuators, i.e. complex control of movements and forces in technical systems, the increasing merging of reality and virtuality allows newer systems to have access to their environments through easier-to-use interfaces, such as in smart home systems. It can be assumed that such artificial cognitive systems will increasingly be used in all areas of life in the future and will interact with humans, i.e. with their natural intelligence.
Perspectives
• Summer 2025 Kick-off workshop of the working group
• Online lecture series from autumn 2025: In an online lecture series, we are planning lectures on AI that are specially prepared for an interdisciplinary audience. The aim is to clearly explain the various algorithms and architectures (e.g. deep learning, reinforcement learning, etc.) of today's AI methods without requiring detailed technical knowledge.
• Summer 2026 or 2027 Academy lecture series: probably four lectures from different disciplines on the topics of ‘AI and humans’
• Spring 2026 ‘AI Promptathon’: in cooperation with the University of Hamburg
Background Information
The topic of artificial cognitive systems is of outstanding social importance both in terms of technical applications and possible risks of AI as well as in terms of possible changes in human perceptions and social norms. It can be assumed that artificial cognitive systems will revolutionise many areas of application and professional fields - from the automation of simple activities to the support of complex decision-making processes. This will lead to a reorganisation of job profiles and may bring both opportunities and challenges for employees. Entire occupational groups may disappear if AI takes over certain activities. It can also be assumed that automated assistance systems will become even more pervasive in everyday life.
With the increasing spread of AI and artificial cognitive systems, the need to prepare people for these technologies is therefore growing. Therefore, education systems need to be adapted to foster the skills needed in an AI-driven world. The way in which humans interact with AI will become increasingly important, with the development of fit-for-purpose systems being crucial to ensure effective collaboration between humans and AI.
The use of AI raises numerous ethical questions - with regard to decisions made by AI systems, the transparency of algorithms and the avoidance of discrimination. The ability to think critically when dealing with artificial cognitive systems is therefore also becoming increasingly important. Artificial cognitive systems can also help to better involve people in social, economic and political processes by facilitating access to information and services. At the same time, there is a risk that technological developments will widen the digital divide if certain population groups are not adequately involved in such processes.
The formulation of guidelines and laws that do justice to progress and at the same time utilise the advantages of technology is of decisive importance for the question of how man and machine will interact in the future.
