Two hours a day of screen time vanished without cutting corners
In a small but carefully measured experiment at a Peruvian university, researchers asked whether artificial intelligence might ease one of the quieter burdens of modern academic life — the exhaustion of endless screen time and dense reading that precedes every lesson. Eight instructors, given AI-generated summaries in place of full texts, found their digital fatigue falling by nearly a quarter and their daily screen hours shrinking by almost two. The study is modest in scale, but its question is large: as teaching migrates ever deeper into digital environments, can thoughtfully deployed tools restore some of the human energy that those environments consume?
- University instructors across disciplines are quietly burning out under the weight of digital preparation demands — a fatigue that standard workload conversations rarely name.
- A controlled eight-week trial in Peru found that AI-generated reading summaries cut daily screen time by 122 minutes and reduced measured exhaustion by up to 25%, with near-perfect protocol adherence.
- Faculty who expected to distrust the technology instead rated it 6.2 out of 7 for usefulness — suggesting the adversarial narrative around AI in education may be softer than assumed.
- The gains came without sacrificing depth: instructors engaged more strategically with content rather than skipping it, pointing toward a model of AI as cognitive relief rather than intellectual shortcut.
- Researchers are urging caution — the sample was eight people at one institution over eight weeks, and questions of academic integrity and long-term generalizability remain wide open.
A research team in Latin America posed a deceptively simple question: could an AI tool make teaching less exhausting? Working with eight university instructors at a private institution in Peru, they designed an eight-week experiment in which ChatGPT produced 200-word summaries of assigned readings along with open-ended discussion questions. Instructors alternated between weeks using these AI materials and weeks following their usual preparation routines, giving researchers a direct comparison.
The results held up across every participant regardless of discipline. Digital fatigue scores fell between 22 and 25 percent during AI-assisted weeks, and instructors logged an average of 122 fewer minutes of screen time per day — nearly a 30 percent reduction. Crucially, the researchers found that instructors were not cutting corners; they were engaging with material more strategically, spending less time scrolling and more time thinking about how to teach it. Protocol adherence reached 96 percent, and statistical effect sizes were strong enough to rule out coincidence.
What surprised the researchers may matter as much as the fatigue numbers. When asked to rate ChatGPT on a seven-point scale, faculty scored its usefulness at 6.2 and its ease of use at 6.4, with every instructor rating both dimensions at least a 5. The common assumption that faculty are inherently resistant to AI appeared to dissolve when the tool was framed not as a replacement for academic judgment but as relief from grinding busywork.
The authors are measured in their conclusions. The study was small, single-site, and short. Larger multi-institution trials are needed before anyone can claim these results travel. Academic integrity also looms as an unresolved concern — as AI summaries become more common, institutions must ensure students continue engaging critically with source material rather than inheriting the same shortcuts. Still, the study opens a quieter possibility: that the relationship between educators and artificial intelligence need not be defined by suspicion, and that the real test ahead is whether benefits observed in one Peruvian classroom can survive contact with the full complexity of global higher education.
A team of researchers in Latin America set out to test a simple question: could an AI tool actually make teaching less exhausting? The answer, according to a new study published in Education Sciences, appears to be yes—at least for the eight university instructors who participated in an eight-week experiment at a private Peruvian institution.
The researchers designed their investigation around a practical intervention. Instead of asking instructors to read entire assigned texts during their lesson preparation, they provided AI-generated summaries—each one roughly 200 words—along with three open-ended discussion questions, all produced by ChatGPT. The instructors then alternated between weeks using this AI support and weeks following their normal preparation routine, allowing the researchers to measure the difference directly.
The results were striking. Using a validated scale designed to measure digital fatigue, the team found that exhaustion scores dropped between 22 and 25 percent during the weeks when instructors relied on AI summaries. This improvement held steady across all eight participants, regardless of their discipline. More tangibly, the instructors logged an average of 122 fewer minutes of screen time per day—nearly two hours—when using the ChatGPT materials, a reduction of about 29 percent from their baseline. Importantly, this efficiency gain did not come at the cost of depth. The researchers emphasize that instructors were not skipping content but rather engaging with it more strategically, spending less time scrolling through dense readings and more time thinking about how to teach.
The study's design was rigorous enough to rule out coincidence. Instructors adhered to the intervention protocols 96 percent of the time, with no reported adverse effects. Statistical measures of effect size—the kind that determine whether results are real or random—came back strong and consistent across multiple cycles, suggesting the fatigue reduction was genuinely tied to the AI tool, not some other factor.
Perhaps equally important was what happened when the researchers asked the instructors themselves whether they would actually use this technology. On a seven-point scale, faculty rated the usefulness of ChatGPT at 6.2 out of 7 and its ease of use at 6.4 out of 7. Every single instructor scored both measures at least a 5, indicating broad acceptance. This finding cuts against a common assumption in higher education—that faculty members are inherently skeptical of AI. The researchers suggest that skepticism may soften when AI is presented not as a replacement for academic judgment but as a tool that reduces busywork and frees up mental energy for actual teaching.
The authors are careful about their claims. The study was small, conducted at a single institution, and lasted only eight weeks. To know whether these results hold up in different settings, with different faculty, over longer periods, researchers would need to run larger trials across multiple universities. There is also the matter of academic integrity—the researchers stress that as AI tools become more common in education, institutions must remain transparent about how they are being used and ensure that students still engage critically with source material rather than relying solely on AI-generated summaries.
What the study does suggest is that the relationship between faculty and artificial intelligence need not be adversarial or suspicious. When positioned as a practical solution to a real problem—the grinding fatigue of managing heavy teaching loads in digital environments—AI appears to be something instructors are willing to embrace. The next question is whether that acceptance will hold as the technology becomes more widespread, and whether the benefits observed in this small Peruvian classroom will translate to the much larger and more complex landscape of global higher education.
Citações Notáveis
When AI is framed as a practical, workload-reducing tool rather than a replacement for academic expertise, it is readily embraced— Study authors
A Conversa do Hearth Outra perspectiva sobre a história
So eight instructors, eight weeks, and suddenly digital fatigue drops by a quarter. That seems almost too clean. What were they actually doing differently?
They weren't doing less work—they were doing different work. Instead of spending an hour reading a dense academic paper, they got a 200-word summary plus three discussion questions, all from ChatGPT. That freed them to think about pedagogy rather than just absorb content.
But couldn't they have just assigned shorter readings to begin with? Why does it have to be AI?
That's fair. But the point is that someone still had to do the summarizing and question-writing. Before, that fell on the instructor. Now it's automated. The time saved isn't trivial—two hours a day per person.
Two hours sounds like a lot. Were they actually cutting corners, or is this real?
The researchers measured it carefully. Screen time dropped 122 minutes on average. And they found no drop in what the instructors actually taught or how students engaged. It was efficiency, not compromise.
What surprised me most was that every instructor rated the tool at least a 5 out of 7. I'd expect more resistance in academia.
That's the key insight. When faculty see AI as a tool that gives them back time and reduces stress, they're not defensive about it. The resistance comes when it feels like a threat to expertise. Frame it as a workload reducer, and the conversation changes entirely.
So what's the catch? Why isn't every university doing this?
Scale and proof. This was one university, eight people, eight weeks. Before institutions invest in rolling this out, they need to know it works across different disciplines, different student populations, different countries. And they need guardrails around academic integrity.