The friction paradox: generative AI and the educational value of difficulty in education abroad
This paper argues that while generative AI can broaden access to education abroad, it risks undermining the core intercultural learning outcomes by mediating the very difficulties that foster independence and perspective-taking, necessitating a shift in focus from banning AI to managing the degree of interpretive agency students retain.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
For decades, the promise of studying in a foreign country has been rooted in a simple, powerful idea: that being surrounded by the unfamiliar forces you to grow. When a student steps off a plane into a culture where the rules of conversation, the rhythm of daily life, and the unspoken social cues are different, they are expected to stumble, to pause, and to figure things out. This struggle is not seen as a failure of the system, but as the engine of learning. Psychologists have long understood that the brain learns best when it encounters "desirable difficulties"—challenges that slow down immediate progress but build deep, lasting understanding. Similarly, the theory of transformative learning suggests that true change happens when our old ways of seeing the world break down, forcing us to rebuild them. For a long time, educators believed that the only way to get this growth was to remove the safety nets, letting students navigate the confusion on their own.
Today, that landscape has shifted. A new layer of technology has arrived, one that does not just provide information but actively interprets the world for the user. Generative artificial intelligence can translate a landlord's angry text, explain why a classmate's silence might mean disagreement rather than rudeness, and draft a polite email to a professor in seconds. This technology sits right inside the cultural encounter, offering a way to bypass the very confusion that was once considered the point of the trip. The question is no longer whether students can survive abroad, but whether they are still learning how to live there.
This paper explores a tension at the heart of modern education abroad, proposing that while artificial intelligence can make studying overseas more accessible, it may also quietly strip away the very mechanism that makes the experience educational. The author, drawing on a background in counseling students for international study, argues that we are facing a "friction paradox." On one hand, these tools can remove harmful barriers, such as the inability to read a medical warning or the panic of not understanding a bus schedule, allowing more students to participate. On the other hand, by smoothing over every bump in the road, the technology risks turning a transformative journey into a frictionless tour, where the student remains physically present but intellectually insulated.
The core of the argument rests on a distinction between different kinds of difficulty. Not all struggles are good for learning. Some difficulties are "harmful friction," like a language barrier that prevents a student from getting emergency medical care or navigating a bureaucratic system; these are obstacles that should be removed. Other difficulties are "productive friction," such as the awkwardness of trying to make a friend in a new language or the confusion of a misunderstood gesture; these are the moments where learning happens, where the student must stretch their understanding and adapt. The paper suggests that artificial intelligence is currently blurring the line between the two. It can easily remove the harmful barriers, which is a good thing, but it often removes the productive ones too, doing the thinking for the student before they have a chance to do it themselves.
The author introduces a concept called the "mediation gradient" to explain why the amount of time a student spends using an app matters less than how they use it. Imagine a student who uses a tool to prepare for a conversation, practicing what they might say and then going into the room to speak for themselves. This is a healthy use of technology that supports learning. Contrast this with a student who uses the tool to write the entire message for them, or to get an instant explanation for a stranger's behavior without ever asking the person directly. In this second case, the student has handed over their "interpretive agency"—their power to make sense of the world—to the machine. The paper suggests that when students hand over this power, they may adapt to their new environment more quickly and feel more comfortable, but they are not necessarily developing the deep intercultural competence that comes from wrestling with ambiguity.
A key finding of the paper is that the current research on international students and artificial intelligence is showing a confusing picture. Recent studies suggest that students who use these tools report feeling more adapted and less anxious. The author argues that this is likely true, but it is measuring the wrong thing. A student can be very well adapted—meaning they are happy, safe, and getting their daily tasks done—while failing to learn how to interpret different cultural perspectives. The tools make life easier, but they might also make the learning invisible. The paper warns that if institutions focus only on student satisfaction and smooth adaptation, they might miss the fact that the educational value of the trip has diminished.
To address this, the paper proposes a new way of thinking about how students interact with technology abroad. It suggests that educators should not simply ban these tools or allow them without question. Instead, they should teach students to recognize when they are using the tool to prepare for a human interaction versus when they are using it to replace one. The goal is "supported immersion," where technology removes the dangerous or exclusionary barriers but leaves the challenging, growth-inducing moments intact. For example, a student might use a tool to learn the vocabulary needed to ask a question, but then they must ask the question themselves and sit with the uncertainty of the answer.
The paper concludes by calling for a shift in how we design these programs. It suggests that the future of education abroad will not be a return to a time without technology, nor will it be a future where technology does everything for us. It will be a middle path where we carefully decide which difficulties to remove and which to preserve. The author argues that we must stop treating all obstacles as things to be eliminated. Some obstacles are the very things that teach us how to be global citizens. If we smooth the path too much, we may find that students have traveled the world, but they have never truly encountered it. The challenge for the future is to build systems that use artificial intelligence to open doors, not to build a wall of convenience that keeps students from ever having to knock on them.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.