Virtual Round Table "ChatGPT & Co."

Daniela Musial-Lemberg

Our speakers Reinhard Heil from the Institute of Technology Assessment and Systems Analysis (ITAS); Jonas Tischer, Senior Software Engineer from VALIDAITOR (KIT spin-off) and Michael Sowka, an experienced IT freelancer, presented ChatGPT from different perspectives. The discussions focused on concrete usage tips for job and everyday life as well as potential dangers on the technical and social level. Important topics that are relevant for both students and entrepreneurs.

We have summarized the most important impulses for you:


ChatGPT is especially good for applicable communication purposes, social media, explanations, recommendations and programming.

As an entrepreneur or freelancer, it can save you a lot of time and optimize your own work processes.

“Creativity does not come at the push of a button”, but ChatGPT can deliver creative texts for various purposes quickly and at any time. (M. Sowka)

At the moment, it is a “balancing act in AI regulation”.

The EU Parliament is deliberating on the new law that will regulate the use of AI in the future.

If you work for a large company, make sure that the use of ChatGPT is desired or allowed.

The important distinction between language models and knowledge models is lost – “ChatGPT as pars pro toto for all chatbot applications based on large language models.

  • ChatGPT can produce plausible-sounding falsehoods – we are talking about so-called AI hallucination. In the field of AI research, it is said that AI hallucinates when chatbots and artificial intelligences, “confidently” give answers that, however, do not seem to be justified by training data. This is because the data used is either insufficient, biased or too specialized.
  • Preparation is key: “Bad request – bad answer” –> you need some practice using ChatGPT.
  • Be especially careful when dealing with customer data, invoices and company secrets. ChatGPT does not inform its users about the way the data they provide is processed. It is also important to keep in mind that the chatbot ChatGPT belongs to the US company OpenAI.
  • In general, use AI if you can validate the results yourself.


The interest of users in AI applications such as ChatGPT is growing. Awareness of the exact potential uses, but also the dangers, needs to be raised in society.


StartUpSecure KASTEL would also like to make a contribution to this and bring the topic back into focus in the fall. At the next Virtual Round Table – which will take place live this time – we will therefore address the legal side and data protection in the use of AI.


We look forward to more exciting round tables – virtual or live!


Do you want to see more?

More blog post