Three promising fields of application for generative AI

9 May 2023

Professor Frank Steinicke, University of Hamburg, outlines risks and opportunities of generative AI

The hype surrounding Chat GPT has conjured up the impression that generative artificial intelligence (AI) can change anything and everything. However, the technology holds both risks and unimaginable opportunities. More than 1,000 researchers and technology experts, including Elon Musk, recently called for a pause in the development of new AI models. Hamburg News spoke to Prof. Dr. Frank Steinicke, Human-Computer Interaction at the Department of Informatics at the University of Hamburg, about three particularly promising fields for generative AI applications.

Professor Steinicke, are modern AI systems dangerous and should the development of the technology be halted?

"I think a stop would be counterproductive because such a ban would not be enforceable globally. Countries that uphold it would fall behind technologically. I would prefer to see a more responsible approach to the development and use of AI involving all relevant stakeholders including ethicists, communication experts as well as computer scientists. The power attributed to AI systems is exaggerated at the moment. We are a long way from superhuman intelligence or so-called strong AI.

The human brain draws on 250,000 years of evolution to process information ideally. Chat GPT draws on vast amounts of knowledge on the internet Which is better?

"Clearly the brain. Chat GPT does not produce knowledge, it is a speech flow model that simulates knowledge and does so excellently but without understanding the solutions it presents. Yet, it is a great tool that can relieve people of tasks like simple coding or texting. So it can replace bad programmers, lawyers or doctors - but never good ones."

Professor Frank Steinicke, University of Hamburg

Fields of application with huge potential for AI systems:

1) Text and image production

"AI systems open up new possibilities, especially for fast and easy production of multimedia content. Start-ups such as Stability.AI, Dall-E, Midjourney, Nightcafe or Pixelz should be mentioned here as they are creating convincing images based on texts or sketches. Trained professionals were needed to set up a personal website in the past. That can now be done in a short space of time. Generative AI is no substitute for a top media designer, but founders can get started quickly thanks to the technology.

2) Medicine

"AI systems are improving from one generation to the next e.g., when it comes to recognising patterns in thousands of X-ray images. That improves the diagnosis, the development of medication and personalised medicine. AI must remain purely a tool for healthcare professionals. Medical decisions will still be made by human beings. We should also keep an eye on international competition when it comes to the limits of AI. Regulations on basic data protection are not as stringent in e.g., China or the United States. Yet, such regulations inhibit technological developments in Germany to an extent.

3) Education

Generative AI is getting better at producing texts for pupils and students. Banning it would not make sense, and they should learn to deal with both artificially-generated knowledge and the technology itself. The child or student's intellectual performance can be graded. AI-supported apps like Anton, Duolingo or Quizlet can also relieve teachers. Instead of "spamming" 25 children head-on with mathematical formulas or history, each student can learn accurately. Then teachers can encourage them to think independently, reflect and discuss." Deep Fake should also be addressed in the classroom, Steinicke said. "Creating plausible but false information with AI and then spreading it on all kinds of social media channels is becoming far easier." Thus, students should be sensitised to it early and encouraged to query sources and to identify signs of manipulation. "If you know how the technology works, you can't be fooled so easily.

ys/sb/pb