As part of a series of workshops designed to help journalism students develop resilience and mental health coping strategies, we led a session where participants used artificial intelligence to walk through a simulation of a traumatic reporting scenario.

We gave ChatGPT the scenario (a stadium collapse at a non-league football match) and instructed it to run a “choose your own adventure” simulation, whereby the students had to tell the AI what they would do at each stage.

They had to suggest practical actions (such as “go talk to the paramedics”), ask questions in interviews, and react to new information they were given.

This session was designed to mitigate students’ anxieties around this type of disaster reporting by alleviating their fear of the unknown. The AI scenario would guide them through an interactive, blow-by-blow narrative, allowing them to consider what would happen and, crucially, what they would do at each juncture.

This would theoretically give them an idea of what to expect in a real-world environment, increasing their confidence to tackle such a scenario.

We found this to be an engaging session. Feedback was positive; students indicated that the prospect of reporting from a potentially distressing news event felt less daunting, having “practised” one in a safe environment.

We did find the AI wasn’t always realistic in its version of events, and didn’t necessarily give the students a faithful representation of what they could expect in the real world.

As a pedagogical tool, the AI simulation fostered active participation and captivated the students. We retain a steadfast belief in its capacity to facilitate experiential learning.

But were we to repeat the session, we would spend more time critically evaluating what was happening in the scenario, beginning a meaningful dialog by analysing how close it was to real life.