In his keynote address at the inaugural NAS Research Conference on 1 July 2025, Prof Michael von Maltitz delivered a wide-ranging and compelling critique of the current state of science education. Speaking to an audience of researchers and academics, he challenged assumptions about learning, assessment, and the role of artificial intelligence (AI) in higher education – offering both caution and practical guidance.

Prof Von Maltitz – from the Department of Mathematical Statistics and Actuarial Science at the University of the Free State (UFS) – opened with an overview of the industrial revolutions leading up to the current Fourth Industrial Revolution, characterised by artificial intelligence, connectivity, and data-driven automation. He warned against remaining entrenched in this phase of development, arguing that AI, while powerful, is not truly intelligent. “AI … is … artificial,” he said. “It is based on brute-forcing very large numbers of very basic operations at blazing speeds, linking external inputs to stored information. And so, it’s not intelligent. It’s just strong.”

He cautioned that the unchecked use of AI – driven by efficiency, not understanding – risks entrenching systems that prioritise ease and profit over education and well-being. “Everything is profit-driven at the moment. Everything, and I mean … everything. Really. It is this greed that keeps us firmly stuck in the Fourth Industrial Revolution.”

This, he suggested, makes the vision of a Fifth Industrial Revolution both necessary and urgent. The next phase, he argued, should be one that centres on sustainability, equity, human-machine collaboration – and critically – the development of human intelligence and critical thinking. “There should be something here about ‘building human intelligence’ or ‘critical thinking’. This would truly make the Fifth Industrial Revolution about bettering humanity.”

When the measure becomes the mission

Central to his address was the idea of ‘broken proxies’ – the phenomenon where a measurement designed to approximate a goal becomes the goal itself, distorting the original purpose. He illustrated this concept using examples ranging from GDP and crime statistics to social media algorithms, before turning to science education. Here, grades and degrees, once indicators of knowledge and progress, have become ends in themselves.

“The only things that are important to students are grades and degrees, because the incentives are linked to grades and degrees, and so, obviously, all effort will go towards grades and degrees.”

Prof Von Maltitz reflected on his own academic journey, describing how he excelled at exams and accumulated qualifications, yet absorbed little meaningful knowledge in the process. “I played the grades game, and nothing stuck in long-term memory, as is the case with many of our students today,” he said. “Why? Well, there were merit bursaries, degrees, and awards up for offer, not for learning, but for performing well.”

This system, he argued, incentivises performance over understanding and leaves students vulnerable to shortcuts – particularly through generative AI. “Under the assumption that rewards are linked to grades and not education, if you offer a student an assessment method that can be gamed … it will be gamed.”

Referencing a recent MIT study, he warned of the cognitive toll of over-reliance on AI. “They showed that, over four months, the AI users’ brains became systematically less active, especially when asked at the end of the study to do a brain-only essay. They had lower brain function in every area. In four months, they had become significantly ‘dumber’ than their counterparts in the other arms of the study.”

Rebuilding curiosity and competence

Despite this sobering analysis, the address was not without optimism. Prof Von Maltitz urged delegates to reimagine education by shifting away from content-heavy teaching and rigid assessment structures. He called for a renewed focus on curiosity, conscious incompetence, and lifelong learning. “Are our students able to self-assess, identify weaknesses and gaps in their knowledge bases, seek answers, and build their own learning paths? Are they humble enough to say, ‘I don’t know’, and curious enough to go and find the answers?”

To support this vision, he proposed four practical steps: redefining teaching goals, distilling module content to its essentials, focusing on graduate attributes such as critical thinking and communication, and reassessing how learning is measured. He encouraged alternatives to traditional exams, including portfolios, interviews, peer assessment, and real-world problem solving.

“We don’t have to pretend to teach students everything in a particular field – but rather we show them what is out there to be learned,” he said.

“Education should not be about teaching everything,” he concluded, “but about showing students what can be known, how to learn, and where to go next.”

About Prof Von Maltitz

Prof Von Maltitz is Associate Professor in the UFS Department of Mathematical Statistics and Actuarial Science. He has a long-standing connection with the university, having been a student at the UFS since the start of his BSc, which he completed with distinction in 2003. Over the following years, he obtained a BCom Honours in 2004, MCom in Economics in 2005, BSc Honours in Mathematical Statistics in 2006, MSc in Mathematical Statistics in 2007, and completed his PhD in 2015 while already lecturing.