What Happens If AI Makes Things Too Easy for Us?

Most people who regularly use AI tools would say they’re making their lives easier. The technology promises to streamline and take over tasks both professionally and personally—whether that’s summarizing documents, drafting deliverables, generating code, or even offering emotional support. But researchers are concerned AI is making some tasks too easy, and that this will come with unexpected costs. In a commentary titled Against Frictionless AI, published in Communications Psychology on 24 February, psychologists from the University of Toronto discuss what might be lost when AI removes too much effort from human activities. Their argument centers on the idea that friction—difficulty, struggle, and even discomfort—plays an important role in learning, motivation, and meaning. Psychological research has long shown that effortful engagement can deepen understanding and strengthen memory, sometimes described as “desirable difficulties.” The authors worry that AI systems capable of instantly producing polished answers or highly responsive conversation may bypass these processes of learning and motivation. By prioritizing outcomes over effort, AI could weaken the experiences that help people develop skills, build relationships, and find meaning in their work. IEEE Spectrum spoke with the paper’s lead author, Emily Zohar, an experimental psychology Ph.D. student, about why she and her coauthors (psychologists Paul Bloom and Michael Inzlicht) argue that friction matters—and what a more human-centered approach to AI design could look like. When you say “friction,” what do you mean, from both a cognitive and an interpersonal standpoint? Zohar: We define friction as any difficulty encountered during goal pursuit. In the context of work, it involves mental effort—rumination and persistence, staying on a problem for some time, and this helps solidify the idea and the creative process. In relationships, friction involves disagreement, compromise, misunderstanding, a back and forth that is natural where you don’t always see eye to eye, and it helps you broaden your horizons. Even the feeling of loneliness is important. It motivates you to find social interactions. So having these negative feelings and difficulty is important in the social context. Given that definition, what do you mean by “frictionless” AI? Zohar: Frictionless AI refers to the excessive removal of effort from cognitive and social tasks. With AI, as we typically use it, it’s really easy to go from ideation right to the end product. You ask AI to solve something with one prompt, and it completes the whole thing. This is a problem because it takes away the intermediate steps that really drive motivation and learning, and it prioritizes outcome over process. Rather than working through the steps, AI does that meaningful work for you. There’s a lot of research showing work products are better with AI. That makes sense, it has all this knowledge, but it does worry us as it may be eroding something essential that will have long-term consequences. If you’re faced with the same problem and AI is removed, you don’t have the required knowledge to know how to face the problem next time. You argue that removing friction can harm learning and relationships. What role do effort and struggle play in human development? Zohar: In learning, the term is “desirable difficulties.” It’s the idea of effort and work, not just any effort but manageable effort. Facing problems that you can overcome, but you have to work at them a bit, that’s the key idea of friction. We don’t want you to face insurmountable problems. We want you to work hard, but still be able to overcome it. This helps you really digest information and learn from it. In interpersonal relationships, you have to face some difficulties to see other perspectives and learn from them, and learn to be accepting of others. If you’re used to an AI reinforcing all your ideas and being sycophantic, you’ll come into the real world and you won’t be used to seeing other ideas. You won’t know how to interact socially because you’ll expect people to always be on your side and agree with you. You won’t learn that life doesn’t always go exactly how you expect it to, and conversations don’t always go the way you want them to. AI’s Impact on Creative Processes A lot of technologies have historically aimed to reduce effort: calculators, washing machines, spellcheck. What’s different about AI? Zohar: Past technologies have mostly focused on reducing physical effort. We don’t have to go down to the lake to wash our laundry anymore. [Past technologies] took away the mundane tasks that weren’t driving our learning and growth, they were just adding unneeded obstacles and taking away time from more important tasks. But AI is taking away effort from creative and cognitive processes that drive meaning, motivation, and learning. That’s a key difference, because it’s not taking away friction from tasks that don’t serve us. It’s taking away friction from experiences that are really important and integral to our development. Are there contexts where AI is already removing beneficial friction? How might the impacts of reduced friction show up over time? Zohar: One clear example is writing. People increasingly rely on AI to draft everything from emails to essays, removing many instances of beneficial friction. Research shows that people trust responses less when they learn they were written by AI, judge AI-generated products as less creative and less valuable, and have greater difficulty remembering their own work products when they were produced with AI assistance. Outsourcing writing to AI strips away both social and cognitive friction. Vibe coding is another good example. If you’re a programmer, coding is integral to what drives your meaning. People get meaning out of their work, and if you’re substituting that with AI, it could be detrimental. The negative impact of frictionless AI is that it takes away friction from things that are really important to who you are as a person, and your skills. One area I worry about a lot is adolescents using AI in general. It’s a really important developmental period to learn and grow and find the path you’ll follow. So if you don’t have these effortful interactions with work and relationships that teach you how to think, this will have long-term detrimental impacts. They might not be able to think critically in the same way, because they never had to before. If they’re turning to AI for social relationships at such a young age, that could really erode important skills they should be learning at that age. What is productive friction? Zohar: Friction goes along a continuum. With too little friction, you’re not getting learning and motivation. Too much friction and the task becomes overwhelming. Productive friction falls right in the middle, where struggle leads to achievement. It’s effortful but possible, and it requires you to think critically and work on a problem for some time or face some difficulty in the process. An example we used in the paper is the difference between taking a chairlift and hiking up a mountain. They both get to the top, but with the chairlift, you don’t get any growth benefits, while the hiker’s climb involves difficulties and a sense of achievement. It becomes much more of an experience and a learning opportunity versus the person who just went up the chairlift effortlessly. Do you envision AI that sometimes deliberately slows people down or asks them to do part of the work themselves? Zohar: It’s important in behavioral science to think about the default option, because people don’t usually change their default. So right now, the default in AI is to give you your answer and probe you to keep going down the rabbit hole. But I think we could think about AI in a different way. Maybe we can make the default more constructive. Instead of just jumping to the answer, it’s more of a process model where it helps you think about the problem and teaches you along the way, so it’s more collaborative rather than a one-stop shop for the answer. How might users of these systems and the companies developing them feel about such a design shift? Zohar: For the makers of these systems, the biggest concern is the pushback. People are used to going in and just getting the answer, and they might be really resistant to a design that makes them work more for it. But it might feed more engagement, because you have to go back and forth and find the answer together. Ultimately I think it has to come from the companies making these models, if they think [a more friction-full design] would help people. Friction-full AI is more of a long-term product. It’s hard to say if that would motivate companies to change their models to include moderate friction. But in the long term, I think this would be beneficial.