As artificial intelligence (AI) becomes increasingly integrated into education and research, debates are intensifying over its impact on learning and intellectual rigor. Recent developments highlight the growing role AI plays in problem-solving, as well as concerns about its influence on traditional thinking and understanding.

In Malaysia, for instance, the latest Programme for International Student Assessment (PISA) results revealed that nearly two-thirds of students struggle to apply mathematics in everyday contexts. However, the report also noted that more than half of Malaysian students use AI tools weekly to assist with their learning, and they generally demonstrate stronger skills in computer-based problem-solving than their overall performance in core subjects might suggest. This indicates that while fundamental skills may need improvement, proficiency with digital tools could offer an advantage as AI becomes pervasive in various fields.

A recent example emphasizing AI’s potential occurred on September 8, when OpenAI announced it had purportedly discovered a singularity solution to the three-dimensional Navier-Stokes equations—one of the unsolved Millennium Prize Problems in mathematics, offering a US$1 million reward if verified. The solution involved approximately 10,000 autonomous AI agents working collectively for 88 hours, followed by a 17-hour verification process by a separate AI model.

OpenAI’s announcement sparked controversy within the mathematics community due to allegations from American mathematicians Tristan Buckmaster and Levent Alpoge, who had been progressing on related problems and claimed that their interactions with OpenAI’s tools were incorporated into the AI’s training data without their consent. This situation has fueled discussion over the extent to which AI is genuinely generating new mathematical insights versus building upon human researchers' prolonged efforts.

Prominent mathematician Terence Tao, a Fields Medal recipient, has been an advocate for human-machine collaboration in advancing mathematics but also cautions against overreliance on AI. Tao emphasizes the importance of the intellectual journey involved in problem-solving, warning that AI shortcuts may deprive researchers of crucial understanding. He criticizes the rush by AI companies to release results primarily for public relations, often leaving human experts the challenging task of validating and interpreting the findings.

This concern extends beyond mathematics. Thomas Dietterich, editor-in-chief of the arXiv research platform, notes a growing trend of submissions by authors who appear not to fully understand their own AI-generated results. Such developments raise questions about the balance between leveraging AI for efficiency and maintaining deep comprehension.

While some fear that AI could erode fundamental skills or displace human thinkers, examples from countries like Singapore and South Korea demonstrate efforts to integrate digital learning alongside critical skill development. Experts suggest that the key lies in fostering both technological proficiency and conceptual understanding to prepare future generations effectively.

Ultimately, the evolving relationship between AI and human cognition challenges educators and researchers alike to embrace technology responsibly, ensuring that the benefits of innovation do not come at the expense of mastering the processes behind knowledge creation.