Beyond Efficiency: Critical Thinking, Trust, and Learning Outcomes in Human-GenAI Partnerships within UAE Higher Education
This study of UAE undergraduate business students reveals that while GenAI enhances engagement, productivity, and trust, it fails to significantly improve critical thinking and may even induce procrastination, highlighting a critical disconnect between AI-driven efficiency and deep learning outcomes.
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
In the modern classroom, a new kind of partner has arrived. It is not a teacher, nor a fellow student, but a machine capable of writing essays, solving problems, and generating ideas in seconds. This technology, known as generative artificial intelligence, has spread through universities with startling speed, changing how students approach their work. For decades, educators have relied on theories about how the mind learns, such as the idea that our brains have a limited amount of mental energy to process new information at any one time. When a tool helps reduce the heavy lifting of basic tasks, it was hoped that this freed-up energy would allow students to think more deeply, reason better, and learn more effectively. The question that now hangs over universities is simple yet profound: does this powerful new assistant actually make students smarter, or does it simply make them faster?
A team of researchers at a university in Dubai set out to answer this question by watching how business students actually used these tools in their daily studies. They did not just ask students what they thought; they tracked how often students used the technology, how they interacted with it, and what they felt they learned as a result. The study focused on a specific group of undergraduate students who were explicitly allowed to use these artificial intelligence tools for everything from brainstorming ideas to drafting final project reports. The researchers wanted to see if using the tool more often led to better grades, deeper understanding, or sharper critical thinking skills. They also looked for hidden behaviors, such as whether the availability of a quick answer caused students to delay starting their work.
The results revealed a story of two different paths. On one hand, the students who used the artificial intelligence tools more frequently and deeply became much more engaged with their coursework. They reported feeling more productive, saving significant amounts of time, and feeling more creative in how they approached problems. The tools seemed to work exactly as intended for these tasks, acting as a reliable engine for getting work done. However, the researchers discovered a surprising disconnect when they looked at the most important measure of deep learning: critical thinking. While the tools helped students finish tasks faster and with more ideas, the study found that simply using the technology did not make students better at analyzing information or thinking critically. In fact, the data showed that the ability to think critically remained largely independent of how much a student used the artificial intelligence.
This finding challenges a common hope that technology naturally upgrades our thinking skills. The study suggests that while artificial intelligence can handle the heavy lifting of gathering information and drafting text, it does not automatically teach a student how to evaluate that information. The students in the study were not passive; they were active. A large majority of them, nearly 90 percent, checked the facts and references provided by the machine before using them. Yet, this careful checking did not translate into a measurable boost in their critical thinking scores. The researchers found that critical thinking was the strongest predictor of a student's overall learning success, but it was not a skill that the artificial intelligence tool itself could generate. Instead, it appeared to be a skill the students already possessed or developed through other means, which they then applied while using the tool.
Another unexpected behavior emerged from the data, one that the researchers describe as a delay in starting work. Because the students knew they could rely on the artificial intelligence to help them finish tasks quickly, they tended to wait longer before beginning their assignments. One group of students, who used the tools heavily, started their projects about a week and a half later than they might have otherwise. This suggests that the efficiency gained from the technology came with a behavioral cost. The speed of the tool created a sense of urgency that was less immediate, leading students to procrastinate. They traded the time saved during the work for time lost at the start, a trade-off that might not be obvious until the deadline approaches.
The study also highlighted a gap between what students did and what they reported. While most students were careful to verify the information the machine gave them, very few formally acknowledged the use of the tool in their final submissions. Only about one-third of the students cited the artificial intelligence in their work, even though they had checked the accuracy of its output. This creates a complex situation where students are acting responsibly by checking facts but are not following standard academic rules about attribution. The researchers noted that this disconnect between verification and citation is a growing concern in higher education, pointing to a need for clearer guidelines on how to use these tools ethically.
Ultimately, the research paints a picture of a tool that is excellent for efficiency but neutral for deep cognitive growth. The artificial intelligence helped students work faster, generate more ideas, and feel more confident in their ability to complete tasks. It did not, however, make them better critical thinkers on its own. The study concludes that for students to truly benefit from this technology, they must bring their own analytical skills to the interaction. The machine can provide the raw material, but the human mind must still do the heavy lifting of evaluation and judgment. Without that active, critical engagement, the speed of the tool may simply lead to faster, but shallower, learning. The path forward for education, the researchers suggest, lies not in banning the tool or assuming it will teach students everything, but in designing lessons that require students to use the tool while simultaneously practicing the difficult work of thinking for themselves.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.