People have mixed feelings about artificial intelligence (AI), depending on how it’s used. A new study suggests people aren’t fully excited or completely against AI. Instead, they judge it based on its usefulness in specific situations. For example, if an AI tool accurately predicts stock prices, some might feel confident using it. But if a company uses AI to screen resumes during hiring, others might feel uneasy about it. The study, published in Psychological Bulletin, explores why people react this way.
Researchers found that people like AI when they think it’s better than humans at a task and when personalization (tailoring to individual needs) isn’t needed. They dislike AI when it’s less capable or when a personal touch matters. This idea, called the Capability–Personalization Framework, comes from analyzing 163 earlier studies with over 82,000 reactions across 93 situations, like using AI for cancer diagnoses. The analysis showed this framework helps explain people’s preferences.
Factors shaping AI acceptance
People prefer AI for tasks like detecting fraud or sorting large datasets, where AI is faster and personalization isn’t key. However, they resist AI in areas like therapy, job interviews, or medical diagnoses, where they want a human to understand their unique situation. This stems from a desire to be seen as special, and AI is often viewed as impersonal, even if it uses lots of data.
Other factors also play a role. People like tangible robots more than intangible algorithms. Economic conditions matter too - those in countries with low unemployment are more open to AI, while those fearing job loss are less comfortable. The study suggests context is everything in deciding whether AI is welcomed or avoided.
The research, supported by grants from the National Natural Science Foundation of China, offers a way to understand these attitudes. While it’s not the final answer, the Capability-Personalization Framework provides a useful guide. Researchers continue exploring how views on AI evolve, recognizing that capability and personalization are key, though not the only factors, in shaping preferences across various scenarios.