Delete search term


Quick navigation

Main navigation

CAI Forms Competence Centre to Support Users and Explore the Impact of Generative Models

In response to ChatGPT, the game-changing new tool, CAI at ZHAW has formed a Competence Centre “GenerativeAI”. The goals of the centre is to study the potential impact of generative models like ChatGPT on society, to perform research on and with these models, and to inform and support interested parties and address questions or uncertainties they might have.

In the past months ChatGPT has caused a stir. Every day dozens of articles, blog posts, tweets etc. are discussing the capabilities and limitations of the model. ChatGPT is used for writing CV’s, completing exams, generating reports, even creating a new operating system. The range of the reactions it’s attracting is immense: from excitement to curiosity to anxiety. Can it revolutionize every-day life like the internet did? How will it change education? What impact will it have on jobs, writing processes and society in general? In light of these varied responses, the Center for Artificial Intelligence (CAI) at the Zurich University of Applied Sciences (ZHAW) has decided to address the implications of this and other powerful generative models.

To this end, CAI has formed a new task force “GenerativeAI”, dedicated to ChatGPT and generative models in general. The team is composed of experts in machine learning, natural language processing, and computer vision, with deep expertise in generative models and the evaluation of such models. Members of the team have used GPT-3, ChatGPT's predecessor, already for a range of projects, such as a natural language interface for a soccer database, creating images from simple keyword descriptions, and co-writing an award-winning article on large language models. 

The CAI task force is currently conducting different experiments with ChatGPT and addressing a wide range of questions related to its applications in industry, research or personal use, as well as its potential drawbacks. Some examples of the current projects are:

  • Internal Study: How successful would a GPT model be in university exams in various disciplines (law, computer science etc.)
  • Industrial Project: How to generate audit reports with GPT-3
  • Scientific Evaluation: Comparison of ChatGPT with the state-of-the-art systems for automatic summaries’ generation
  • Organization of open discussions among industry and academia representatives.