I am head of the health and healthcare area of ISTUD and an expert in narrative research.
How can AI be applied to narrative research?
In the world of narrative research, Artificial Intelligence (AI) is opening up new avenues, making our work increasingly efficient and innovative. We at ISTUD, for example, use MAXQDA, a state-of-the-art software for qualitative data analysis. At this year’s Berlin conference, organized by Verbi Software (the company behind MAXQDA), the AI Assistant was unveiled, an innovation that promises to transform the way we work.
The AI built into MAXQDA offers two main features: automatic transcription of audio into text and assistance with research, providing suggestions and insights directly from ongoing analyses. This virtual assistant, which incorporates some chatGPT capabilities, helps researchers save time and work with large amounts of data.
However, there is no shortage of challenges. A major concern is privacy. The AI Assistant saves texts in an internal database for up to 60 days, which raises questions about the management of sensitive data. At ISTUD we deal with disease narratives, and until we are sure about privacy management, we prefer not to use this feature.
Importantly, although the AI Assistant can provide valuable suggestions, it cannot completely replace the human eye. Researchers need to compare available information and data, maintaining a central role in the analytical process. In addition, AI can be influenced by biases in the source data, such as cultural, racial, and gender biases.
For example, we notice biases when we use AI to create images for our projects: stereotypes often emerge, such as the representation of the male doctor holding a stethoscope. This shows that AI reflects the biases present in the data it is trained on.
Recently, an Italian AI was developed based on data from the European Union, designed to reflect European culture in its responses. This suggests to me that in the future we will have many AIs, each specializing in data from specific regions, cultures, or policies. Consequently, it will be critical to develop a strong critical sense to use these technologies in an informed and informed way.
In conclusion, AI is changing the landscape of narrative research, offering powerful tools but also requiring careful and informed management by researchers. With the right balance, AI can become a valuable ally, helping us explore new frontiers of knowledge.
But on a day-to-day basis, is AI making your work easier?
AI is facilitating our daily work, although it is in any case one more tool to learn. Tools such as those for creating images, which we would previously have had to get from free databases, and chatGPT, which we use for reviewing emails or getting summaries on various topics, help us speed up many of our tasks.
We often use chatGPT as an additional source of information for research. However, we should be aware of its limitations: chatGPT does not perform a comprehensive literature review and often does not cite the sources or data from which it draws information. This may limit its usefulness in contexts where rigorous source verification is required.
More specific AI software exists for tasks such as writing scientific papers, but so far we have not used them. These tools could offer significant advantages, but it is critical that humans maintain a critical role in evaluating the results produced by AI.
Indeed, human judgment remains irreplaceable. Technology can accelerate many processes, but the interpretation, critical evaluation, and contextual application of information always requires the intervention of an experienced researcher.
Is the use of AI fundamental to you?
For us it is not fundamental and our use of it is still restricted. It has potential to be a valuable aid in speeding up some tasks that at the moment, especially for privacy reasons, are applied little. It is one thing to work with data that is already online, i.e., in open access, and it is another to do it with anonymous narratives from patients, family members and physicians that are covered by privacy because they are very sensitive. However, I think it is something to be monitored and started to be considered because it may take on greater roles in the future.
Earlier we were talking about privacy, and I wonder if the use of AI might somehow complicate your work?
In our projects, GDPR compliance is critical. These European privacy regulations require that if some data is stored on servers outside of Europe, this must be declared to the participants of the storytelling projects. This complicates matters because it is always necessary to communicate to participants all the constraints and how the data will be handled. In some cases, these restrictions are so limiting that they prevent us from entering data online.
Also, people who participate in our projects, telling their narratives, do so anonymously. This is something we focus on, as anonymity gives participants the opportunity to tell their stories more freely and truthfully. However, for some, knowing that their story will be recorded by an AI may not be reassuring and may hinder their willingness to share.
Would you like to add more?
Right now AI is a support. We use it in research still little for privacy reasons and more in other aspects. We need to start understanding it and getting to know it because developments are fast. I am skeptical about thinking it will replace humans in complex or creative work because I also see that most companies using AI are very thorough. At the training level it is useful to teach the younger generation how to use AI but always with critical thinking.
In terms of creativity, will AI be able to replace humans?
I don’t think AI can completely replace human creativity. A practical example comes from the graphics program I use. It is crucial to understand how to formulate questions to the AI, as the result varies greatly depending on how the query is set. The AI needs a lot of cues to produce quality feedback, since it is not capable of being minimalist. The more you ask it for something simple, the more of a challenge it becomes for it. In my opinion, in the future there will be specialists dedicated to formulating the right questions to AI. For example, to generate images, I happened to have to ask for output up to thirty times. Instead, having a professional graphic designer on the team increases the chances of getting the right image on the first try. AI’s main difficulty in meeting human expectations lies in the fact that the results produced often tend to be mainstream and fail to capture the uniqueness and specificity required in a creative project. Human creativity, with its capacity for insight, interpretation and innovation, remains irreplaceable in many areas.
In summary, while AI can be a powerful and useful tool, human creativity retains a unique and irreplaceable value. The ability of a professional to capture nuance, emotion, and specific details, especially in unconventional contexts, is something that AI, at present, cannot fully replicate. Collaboration between AI and human creativity even in research can lead to extraordinary results, but it is unlikely that AI can completely replace human ingenuity.