A new study by researchers at the University of Toronto cautions that while Artificial Intelligence (AI) is making intellectual work and social interaction easier, this convenience may come at a psychological cost. Published in the Communications Psychology, the research argues that AI’s defining strength, its ability to remove friction from daily tasks, could also weaken essential human processes tied to growth and meaning.
The authors suggest that friction, defined as effort, struggle, and difficulty, plays a crucial role in learning, motivation, and the development of meaningful experiences. Unlike earlier technological innovations that eliminated unnecessary obstacles, AI can also strip away beneficial challenges. Psychological evidence shows that effortful engagement enhances memory, strengthens understanding, and builds skills. When AI systems generate instant, polished outputs, they may bypass the cognitive processes that foster competence and confidence.
The study also examines AI’s expanding role in providing emotional support. While AI companions can help reduce feelings of isolation, researchers warn that loneliness serves as an important social signal. It motivates individuals to seek out and invest in real human relationships. Removing this signal through artificial companionship may reduce the drive to build and sustain authentic connections.
However, the researchers acknowledge that AI offers significant advantages in productivity, accessibility, and well-being. Its impact may vary depending on life stage and context. Experienced professionals may benefit from increased efficiency, and socially isolated individuals may find value in AI support. In contrast, younger learners and those in formative stages risk missing out on the developmental benefits of struggle and sustained effort.
The study concludes that the challenge lies not in AI itself, but in how society chooses to integrate it, balancing technological convenience with the preservation of experiences that foster growth, resilience, and meaning.