Dr Angelica Henestrosa successfully completed her doctorate recently with the defence of her thesis on the use of AI in science communication and how it is perceived by readers - thereby addressing a highly topical issue that is becoming increasingly relevant to society.

In her dissertation, the doctoral student from the Knowledge Construction lab focused on the question of how generative artificial intelligence is changing science communication - a topic that could not be more topical in the age of ChatGPT & Co. The focus is on the perception of AI authorship in science communication texts. The key finding: readers hardly differentiate between texts written by humans and those written by AI - at least in terms of their credibility. Although many are generally open to AI authorship, there is often a lack of in-depth knowledge about how it works and its limitations. ‘Many readers did not realise, for example, that the texts or answers of an AI are always based on probabilities and not on real understanding and that factual accuracy - especially in the case of new topics, which is often the case in connection with scientific content - is not always a given,’ explains the researcher.

It is particularly noteworthy that even explicit references to the strengths or weaknesses of the AI only have a minor influence on the perception of the content. From this, Dr Angelica Henestrosa derives important recommendations for the responsible use of generative AI in science communication - with relevance far beyond the academic world.