Urgent.News

What's breaking now, across thousands of outlets.

Editions

AI

AI Boosted Homework Scores, Then Exam Scores Dropped

An anonymous reader quotes a report from The Economist: Students and school children are increasingly using artificial intelligence. A survey last year by Chegg, an ed-tech firm, found that 80% of rich-world undergraduates used it in their studies. More recent polls put the figure at 94% in Britain and 93% in Germany. Such widespread adoption has fueled concerns about the impact the technology…

An anonymous source cites a recent report in The Economist revealing the intriguing, albeit surprising, effects of artificial intelligence (AI) on students' academic performance. The study, conducted by researchers from Stockholm University, the University of Hong Kong, involved tracking the academic progress of 27,000 Chinese students aged between 12 to 18. Around 80% of these students admitted to using AI models like Doubao and DeepSeek, while the remaining 20% served as a control group.

The findings painted a contrasting picture of students' academic outcomes. After six months of utilizing AI, students' average homework scores increased by 18% across all subjects, and the time they spent on each assignment shrank from an average of 64 minutes to just 45 minutes. However, the most jarring revelation came from their exam scores.

The same students who had benefited from AI-enabled homework, scored 20% lower on their exams than their peers who had not used AI for their studies. This paradoxical outcome suggests that while AI can indeed enhance learning productivity, it is not without its risks. The key takeaway from this study, as highlighted by The Economist, is that students must use AI technology wisely, resisting the urge to rely on AI-generated answers without first trying to solve the problem themselves.

Written by urgent.news from Slashdot's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at news.slashdot.org →

More in AI

Nvidia just showed that the harness, not the AI model, is now the real hero

Nvidia research shows that AI agents can perform well, and not go off the deep end, through fine-tuning, even if the AI model isn't that great at the task.

  • Nvidia's research reveals harness, not AI model, is crucial for long-horizon tasks.
  • Optimized harness and supervisor component boost Claude Opus 5's ARC-AGI-3 score to 100%.
  • Harness includes memory, context, feedback, and scaffolding for AI agents.

More from Friday 21 August →