ARGE Conference, Prag, September 2023
"Jürgen Degenfellner (lecturer for statistics at the ZHAW Zurich University of Applied Sciences) demonstrated to the participants in a highly entertaining manner the nearly endless possibilities of artificial intelligence, while also acknowledging potential risks. The hype around ChatGPT has made AI accessible to a wide audience in a short period. Through AI, machines can now analyze colloquially formulated questions, solve mathematical proofs, or even write complex computer programs, rather than just playing chess. However, this is also the beginning of a narrow tightrope walk. For instance, vulnerability analysis of computer programs can serve to secure one’s own systems, but it can also be a gateway to cybercrime. The biggest danger with AI might be that algorithms for problem-solving are not always predictable or comprehensible for users. Degenfellner concluded his guest lecture with a few provocative statements and food for thought. AI might eventually kill all humans, but no one knows if, when, and how exactly that will happen. The internet may become a place that is less trusted due to falsifications and similar issues than during the era of AI systems. Many of the so-called AI tools are still not robust and secure enough to be reliably used in a professional environment. But one thing is certain: these tools will not get worse and already offer enormous potential for productivity gains today."
