For Fat Thursday in Sano there were announced donuts (which is traditional in Poland) and a discussion about ChatGPT (which is because we like to discuss something important then we meet). We don’t know which of the items brought more of our colleagues to the office that day, but it was really nice to see so many enthusiastic faces in one place. 

Conversational AIs in general, and ChatGPT as one of the most famous examples of them, already gained a reputation as both dangerous and positive tools which can significantly change human role in scientific research, education, and all possible creative occupations. Marian Bubak, Sano’s Scientific Affairs Director, presented his observations on the topic and invited colleagues to share their opinion. Although there were many different considerations, the idea everybody agreed on – the research community has to continue debating about the consequences of the application of Conversational AIs and should elaborate on the rules for regulating its usage.  

Some of the questions which still need to be discussed are:  

• Which research tasks should or should not be outsourced to large language models (LLMs)?  
• Which academic skills and characteristics remain essential to researchers?  
• What steps in an AI-assisted research process require human verification?  
• How should research integrity and other policies be changed to address LLMs?  

Articles which were used as a starting point for the conversation: 

 – ChatGPT: five priorities for research  

 – 10 reasons to worry about generative AI