How Different Personalities Interact With Artificial Intelligence

As artificial intelligence (AI) systems become integral in critical fields like healthcare, aviation, and search-and-rescue operations, the need for effective human–AI collaboration grows. In high-stakes environments, professionals can potentially work in partnership with AIs. Importantly, the experts and professionals who are leading out in these situations rely on accurate, tailored recommendations from AIs to make wise decisions across diverse problem situations. A key challenge lies in designing AI recommendation systems to align with individual user preferences and personalities, such that human–AI partnerships and teamwork become increasingly collaborative and effective.

Dynamics of Strategy Recommendation Systems

A recent study by Dodeja and colleagues examined the dynamics of strategy recommendation systems, exploring how their design can enhance usability and user satisfaction. By examining these recommendations in terms of alignment with users’ preferences and personality traits, the researchers aimed to uncover insights that could inform the design of more effective and personalized AI systems.

In their study, Dodeja and colleagues focused on how users perceive and respond to various types of strategy recommendations when playing the game "Risk." Risk is a game in which players take on the role of world leaders trying to conquer territories by moving armies and battling opponents. The main goals are to gain control of as many countries as possible, using strategies like forming alliances and planning attacks. The researchers assessed users' baseline gameplay preferences and then investigated the effects of different strategy recommendation approaches provided by their AI collaborator—namely where the AI suggests one best strategy, a few similar strategy options, a mix of different options, or every strategic option available. The research team then analysed the effect of these AI approaches on the human user experience.

The study also examined how individual differences in personality influence users' interaction with strategy recommendation systems. All 60 participants in the study completed the Mini-IPIP (International Personality Item Pool) questionnaire, which evaluates the five major personality traits: agreeableness, conscientiousness, openness, neuroticism, and extraversion. With the addition of this data, Dodeja and colleagues were able to explore how these traits impact user perception of the strategy recommendation system.

Study Findings

A number of interesting research findings were revealed. First, the degree to which participants felt that recommendations aligned with their inherent gameplay preferences significantly influenced their perceptions of the system. Increased perceived alignment was associated with higher usability ratings and a greater sense of perceived intelligence of the recommendation system.

Second, in general, participants’ dominant strategy recommendation preference involved being presented with the most relevant strategy rather than a longer list of strategy options. This trend in the data suggests that most users found it easier and more satisfying to receive one tailored recommendation that closely aligned with their gameplay style rather than being presented with multiple options that could potentially lead to confusion or decision fatigue. However, individual differences in personality moderated this effect, and, notably, this preference did not hold for individuals with high levels of conscientiousness. Those high in conscientiousness tended to prefer more comprehensive options rather than just a single recommendation, indicating that their approach to decision-making might involve a desire to consider multiple strategies before settling on a course of action.

Also, the perceived workload while using the recommendation system varied across participants depending on personality traits. In particular, participants higher in agreeableness experienced higher workloads—an experience may have arisen from an inclination to consider all recommendation options comprehensively. Overall, the findings suggest that personality can influence preferences for how recommendations are delivered, and personality can also influence the cognitive load experienced when using AI recommendation systems.

The findings from the study provide valuable insights for developing user-centric AI systems in collaborative environments, suggesting the potential need for customization of recommendations based on user preferences and personality types. Tailoring strategies to support individuals and teams working across different problem contexts will be important in efforts to enhance productive teamwork dynamics, user satisfaction, and technology adoption in various domains.

More generally, it’s essential to build trust so people can feel safe using these tools, and, in critical fields like healthcare, aviation, and search-and-rescue operations, it’s important that professionals and experts working in the field retain control over decision-making and the flexible use of transparent and user-friendly AI support tools. In this way, designers remain focused on creating systems that help users find the best strategies and make the best decision given the circumstances, and they also respect the needs and preferences of human users. By focusing on these design considerations, AI technology is designed to be a helpful companion, promoting confidence and satisfaction in our interactions with AI systems.