RHET AI at ‘AI in Higher Education Teaching in Baden-Württemberg’

On March 11th, 2024, Sci­ence Min­is­ter Petra Olschow­ski invited around 150 uni­ver­sity lec­tur­ers from all over Baden-Württem­berg to take part in the net­work­ing event: ‘AI in High­er Edu­ca­tion Teach­ing in Baden-Württem­berg’. The aim of the event, which took place at the Stut­tgart Uni­ver­sity of Cooper­at­ive Edu­ca­tion (DHBW), was to show­case the diversity of AI tools and their poten­tial bene­fits for teach­ing and teach­ing con­cepts, to pro­mote exchange between lec­tur­ers and to provide inspir­a­tion for their own teach­ing and aca­dem­ic practice.

The day con­sisted of two parts: In the morn­ing, research­ers were able to present dur­ing a poster ses­sion and exchange best prac­tice mod­els from their own teach­ing exper­i­ence. In the after­noon, poten­tial teach­ing con­cepts for AI were dis­cussed and tested in an ‘ideas camp’. In between, there was plenty of time for lec­tur­ers to exchange ideas and to net­work. The event ended with a pan­el dis­cus­sion. One spe­cial fea­ture was the exper­i­ment­al attempt to have an AI tool accom­pany the entire event and sum­mar­ise its content.

The RHET AI Cen­ter was rep­res­en­ted at the event by Markus Gott­schling and Salina Weber. As part of a poster ses­sion on good prac­tice mod­els from teach­ing, they presen­ted their work­shop concept: ‘Prompt towards a goal? The rhet­or­ic of gen­er­at­ive AI’.

Dur­ing the work­shops, the par­ti­cipants devel­op rhet­or­ic­al AI skills in six steps that go bey­ond a purely tech­nic­al under­stand­ing of arti­fi­cial intel­li­gence and AI tools. The aim is to provide ini­tial sup­port for (schol­arly) work with gen­er­at­ive AI, to remove usage bar­ri­ers and high­light lim­it­a­tions and oppor­tun­it­ies. The train­ing takes place in aca­dem­ic and non-aca­dem­ic contexts.

The par­ti­cipants gain a num­ber of insights in the course of the work­shop. First of all, AI tools are a good way to sup­port one’s own (sci­entif­ic) work. By using nat­ur­al lan­guage, for example, the res­ults gen­er­ated by AI tools are low-threshold and eas­ily access­ible. How­ever, gen­er­at­ive AI can­not pro­duce suit­able res­ults on its own. They are trained to imit­ate what is known and provide the most likely answer. There­fore, humans are always needed for AI to be func­tion­al. Humans, or the users of AI tools, have a wide range of con­trol options and sig­ni­fic­antly influ­ence the out­put of the AI through their prompts. If users are trained to devel­op an aware­ness of exactly what they want to achieve with their prompt, i.e. if they have a rhet­or­ic­al goal of com­mu­nic­a­tion in mind, the AI can gen­er­ate more appro­pri­ate res­ults. Co-cre­ativ­ity is the core mes­sage of the work­shop. Through con­scious and rhet­or­ic­al co-cre­ativ­ity, AI tools can be used in an ideal way and be a great sup­port for one's own (sci­entif­ic) work. How­ever, it is always import­ant to be aware of the lim­it­a­tions of AI tools and to crit­ic­ally ques­tion the results.

All in all, our RHET AI team is pleased with how the event went. In addi­tion to the many oppor­tun­it­ies for net­work­ing and dia­logue, the present­a­tion of the dif­fer­ent teach­ing mod­els was very excit­ing and — like the day as a whole — inspir­ing and useful.