Do your homework, but do it right
A year ago, I wrote a two-part series about a fascinating research study from MIT that showed that using AI for essay writing makes you feel like you did good work when in truth you didn’t. One problem is that using AI or even Google search changes the activation patterns in your brain
In a recent study from MIT, researchers found that using AI for essay writing can make students feel like they've done well on their assignments when, in reality, they haven't. One consequence of this is that using AI or even conducting a simple Google search can alter the brain's activation patterns, consequently affecting the ability to learn from the activity.
With AI tools now prevalent in schools and universities, it's no longer necessary to depend on restricted lab studies to comprehend how AI utilization impacts studying and knowledge acquisition.
A comprehensive analysis of 26,811 Chinese students in grades 7 to 12 shed light on how generative AI has impacted their learning experience. The study revealed the consequences of using generative AI for homework, the alterations in homework scores, and the changes in exam performance. The results are rather alarming.
The introduction of generative AI led to a significant reduction in the time students spent on their homework. Previously, students devoted between 50 and 80 minutes to their homework assignments. Once AI became accessible, homework time dropped to as little as 25 minutes. It might appear that students now have more time to engage in video games, but the reality is far more concerning.
Although the utilization of generative AI seemed to benefit students in terms of homework scores, the impact is less impressive. Even though time spent on homework was halved, homework scores increased by only around 20%. However, when it truly mattered—in an exam where AI is unavailable and students must rely on their memory of previous classes and the comprehension of what they have learned—students who used generative AI experienced a decline in their exam scores, typically by about 20%.
In essence, students who relied on generative AI outsourced the rigorous mental labor of studying to the machine. While they had access to the AI, their homework scores remained unchanged, or in some cases, they even improved. Nevertheless, once confronted with a situation where they lacked access to the AI or the machine was no longer helpful, the mental outsourcing backfired, and these students ended up performing worse than if they had invested time in studying and completing their homework properly.
Written by urgent.news from Klement on Investing's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
