How people manage their feelings, whether by rethinking a situation, distracting themselves or bottling things up, has a strong bearing on their mental health. Psychologists rely on questionnaires to measure these habits, but building a good one takes time and expert effort. A new study from Saudi Arabia shows how artificial intelligence can help with the job while humans stay firmly in charge.
Researchers at King Abdulaziz University in Jeddah developed a new Emotion Regulation Scale for adults. The findings appear in BMC Psychology. The team used several large language models, the kind of AI behind popular chatbots, to help write questions, refine wording, screen for bias, translate items into Arabic and interpret the results.
Emotion regulation means the ways people influence which emotions they have, when they have them and how they express them. The new scale was built around a well known model that sees regulation as a series of steps, from choosing situations to changing how one responds. Existing questionnaires often focus on a single strategy or on difficulties, while this one aims to cover a broader range of thinking, behaviour and context.
The AI tools produced an initial pool of 25 questions, which psychology experts reviewed for accuracy, clarity and fairness. Between January and March 2026, 743 Saudi adults completed the questionnaire online, and the sample was split in two so the structure could be found in one half and tested in the other. The final version has 17 questions grouped into four areas.
One area captures unhelpful thinking, such as dwelling on negative feelings or seeing setbacks as proof of failure. The others reflect more adaptive strategies. People who scored high on unhelpful thinking reported more depression and anxiety and lower wellbeing and quality of life, while those who used adaptive strategies showed the opposite pattern.
Scores also differed clearly between people with low and high levels of depression and anxiety. Reliability was acceptable for an early stage scale, though below the level usually expected of established tools.
The findings suggest that AI can speed up the slow early stages of questionnaire design without replacing expert judgement and proper testing. The authors stress that theory, human oversight and real world data remained central at every stage.
The study has clear limits. Participants were recruited online through social media, so they may not represent all Saudi adults. Both halves of the sample came from the same group, the structure was partly refined after the first analysis, and the scale has not yet been tested for stability over time.
Artificial intelligence is spreading quickly through scientific research, and psychology is no exception. This study offers an early example of how chatbot technology might help build mental health measures, while showing that the usual checks still matter.
