What role does AI play today and in the future in aviation? And to what extent are AI applications safe enough for the safety-conscious aviation industry? These questions were the focus of a week dedicated to AI at ZAL. In three events, experts presented insights into the current state of research, current use cases, and forecasts for future applications.
AI at Lunch
For many people, the release of ChatGPT this year was their first conscious encounter with AI. However, few users understand what lies behind it or how to “train” AI for their own purposes. An introduction to the topic was provided by Dr. Leonid Lichtenstein, Head and AI expert at ZAL GmbH, during a lunch, and participants were then invited to a live demo in the AI Lab. Here, interested individuals could try out automatic image recognition for themselves.
Industrial AI Applications
The ZAL Discourse presented concrete industrial applications that can be improved through the use of AI. At the same time, it became clear that the use of AI should always be viewed critically and presents new challenges.
Martin Gromniak, AI expert at ZAL GmbH, explained the case of medical delivery drones, highlighting their potential in saving time by flying over traffic jams and reducing emissions or operating with green energy. For safety reasons, the drones need to be capable of quickly identifying an improvised landing site in case of incidents, which they can recommend to their operator. This is a challenge that can be overcome with AI, provided it is safe, which is the case with explainable AI.
Looking beyond, it was evident that the shipping industry is also embracing AI applications, particularly in maintenance processes. Here, AI is expected to be used for faster motor condition detection, thus supporting the work of machinists. This allows for better maintenance planning and helps avoid breakdowns,” explained Michael Windmann, AI researcher at Helmut Schmidt University in Hamburg. However, a challenge lies in transferring the years of expertise from machinists to an AI model, and adapting models trained on one type of engine to other engine types can be difficult.
Another perspective on AI and data security was presented by Michael Welsch, founder of Metric-space.ai. Many companies are already using ChatGPT in their daily work without much concern for data transfer compliance. However, this can be avoided. Additional programs can automatically recognize critical inputs such as passwords, names, and personal data and anonymize them before transmission. Regarding sustainability, Michael Welsch pointed out that resource-intensive AI applications also consume a significant amount of resources. He used the example of a half-liter of water required to cool ChatGPT servers for a single request.
AI in Five Years
The conclusion of the AI week was the ZAL Innovation Talk “AI in Aviation – 5 Years From Now.” In an exciting interview, Lothar Hotz, Chairman of the Artificial Intelligence Center Hamburg (ARIC), shared his insights into future developments.
AI developers are entering an exciting but also challenging period, as perhaps only one in ten projects will succeed. Nevertheless, it is important to start now rather than waiting. He also emphasized that while there is much talk about AI models, not enough attention is given to the data on which they are based, and the data foundation often plays a crucial role in success or failure.
The ZAL AI week provided participants with a realistic insight into the state of artificial intelligence in aviation. The future promises exciting developments and challenges. It is crucial to seize these opportunities now to ensure that AI can be used safely and in a sustainable manner.
Would you like to attend the next ZAL event? Then sign up for our event newsletter.
