Decoding AI Perception: Capability vs. Personalization – Proaitools
Are you excited or hesitant about using AI tools? A recent study reveals that our feelings towards AI are more nuanced than simple enthusiasm or aversion. Instead of universally embracing or rejecting AI, we carefully weigh its practical benefits on a case-by-case basis.
Consider this: if an AI tool accurately predicted your stock investments, would you use it? What if an AI system screened resumes at a company where you were applying for a job? Our comfort level depends on how we perceive AI’s capabilities and the need for personalization.
MIT Professor Jackson Lu, co-author of a new paper in Psychological Bulletin, explains, “AI appreciation occurs when AI is perceived as being more capable than humans and personalization is perceived as being unnecessary in a given decision context. AI aversion occurs when either of these conditions is not met, and AI appreciation occurs only when both conditions are satisfied.”
The study, titled “AI Aversion or Appreciation? A Capability–Personalization Framework and a Meta-Analytic Review,” offers a new framework for understanding our complex attitudes towards AI.
Capability–Personalization Framework Explained
Past research on AI has produced mixed findings. Some studies highlighted “algorithm aversion,” where people were less forgiving of AI errors than human mistakes. Others showed “algorithm appreciation,” with people preferring AI advice over human advice.
To reconcile these conflicting results, Lu and his colleagues analyzed 163 studies comparing preferences for AI versus humans. They tested their Capability–Personalization Framework: the idea that perceived AI capability and the need for personalization determine our AI preferences.
The meta-analysis, encompassing over 82,000 reactions across 93 decision contexts (e.g., AI in cancer diagnoses), confirmed the framework’s validity. People favor AI when they believe it’s more capable than humans and the task doesn’t require personalization.
“Both dimensions are important: Individuals evaluate whether or not AI is more capable than people at a given task, and whether the task calls for personalization. People will prefer AI only if they think the AI is more capable than humans and the task is nonpersonal,” Lu notes. He emphasizes that high perceived capability alone isn’t enough; personalization matters.
For instance, we tend to prefer AI for fraud detection or large dataset sorting—areas where AI excels in speed and scale and personalization is unnecessary. However, we resist AI in therapy, job interviews, or medical diagnoses, where we value human recognition of unique circumstances.
Lu explains, “People have a fundamental desire to see themselves as unique and distinct from other people. AI is often viewed as impersonal and operating in a rote manner. Even if the AI is trained on a wealth of data, people feel AI can’t grasp their personal situations. They want a human recruiter, a human doctor who can see them as distinct from other people.”
Additional Influencing Factors
The study also identified other factors affecting AI preferences. AI appreciation is stronger for tangible robots than intangible algorithms, and more pronounced in countries with lower unemployment.
“If you worry about being replaced by AI, you’re less likely to embrace it,” Lu observes.
Lu continues to investigate our evolving attitudes toward AI. While this meta-analysis isn’t definitive, he hopes the Capability–Personalization Framework will provide valuable insights into how we evaluate AI in various situations.
“We’re not claiming perceived capability and personalization are the only two dimensions that matter, but according to our meta-analysis, these two dimensions capture much of what shapes people’s preferences for AI versus humans across a wide range of studies,” Lu concludes.
