Big Tech experts and university professors explored how AI is accelerating the shift toward skills-based education, emphasising the role of universities in developing essential competencies such as critical thinking and ethical judgment alongside technical proficiency.

In photo (from left to right): Mr Alfonso Asensio (Google, United States), Dr Sean McMinn (The Hong Kong University of Science and Technilogy, Hong Kong), and Dr Matthias Frey (Sony SCL, Japan) during their panel presentation at ACE2025

The panel discussion ‘AI and the Skills Turn: Driver or Disruptor of Skills-Based Education?’ examined the systemic shifts occurring as artificial intelligence redefines the relationship between academic preparation and professional viability. Moderated by Dr Justin Sanders of Woven by Toyota, Japan, the session brought together Mr Alfonso Asensio of Google, United States; Dr Matthias Frey of Sony SCL, Japan; and Dr Sean McMinn of The Hong Kong University of Science and Technology, Hong Kong to examine how AI is reshaping what it means to be ‘educated’. 

The discussion opened with the acknowledgement that the traditional 200-year-old discipline-based degree model is being challenged by a more modular approach to education. Dr Sanders highlighted that nearly 40% of core workforce skills are expected to change within the next few years, necessitating a radical rethinking of curriculum design. The panellists agreed that while the ‘skills turn’ has been discussed for decades, AI acts as a significant catalyst that accelerates the need for immediate institutional adaptation. Akin to Professor Suzuki, they agreed that the focus must move away from static knowledge toward a dynamic ability to acquire and apply new competencies. This transformation is not merely about adding technical tools but about reimagining the very purpose of a university education in a tech-saturated market. The speakers emphasised that institutions failing to pivot towards this competency-driven model risk becoming obsolete in a landscape where employers increasingly value demonstrated abilities over formal credentials.

A major thematic pillar of the discussion was the critical distinction between short-term technical ‘hype’ and long-term cognitive resilience. Dr Frey observed that while there is currently a high market demand for specific skills like prompt engineering, these may soon be automated or rendered by more intuitive AI interfaces. The panel urged educators to focus instead on durable human skills such as critical thinking, fact-checking, communicating ideas, and the ability to verify AI-generated outputs. Dr McMinn introduced the concept of metacognitive awareness, arguing that students must understand their own thinking processes to effectively partner with machine intelligence. There was a shared concern that overrelying on AI for basic tasks might strip away the grit and problem-solving stamina that students traditionally develop through cognitive struggle. Mr Asensio added to the discussion by highlighting the necessity of ‘discernment,’ the high-level ability to judge the boundaries of what an AI can and cannot do effectively. This discernment allows human workers to identify where their unique ‘added value’ is most impactful, ensuring they remain relevant as AI capabilities expand. The panel concluded that the ultimate goal of education should be to produce wise graduates who can navigate the ethical and practical complexities of AI usage.

The conversation then shifted toward the practical integration of AI within institutional frameworks and the management of global technological biases. Dr McMinn detailed the four-domain framework used at his institution, which categorises AI’s impact into domain knowledge, procedural knowledge, technological knowledge, and cognitive skills. This structured approach allows for a more nuanced assessment of how AI changes specific tasks within a discipline without completely replacing the human element. The panellists also addressed the inherent Western-centric bias in major Large Language Models (LLMs), which are largely developed in the United States. Dr Frey suggested that educators should seek out regional models like Singapore’s SEA-LION or local adaptations of open-source models to better understand different biases in LLMs. Dr McMinn also gave an example of how Hong Kong developed its own model based off of Deepseek for Cantonese speakers in the absence of ChatGPT in Hong Kong. Although this localised approach to AI is seen as a way to empower students in non-Western contexts to lead within their local frameworks rather than merely adopting foreign perspectives, cognitive skills are still at the heart of the discussion. Dr McMinn suggested that students should be aware of, and trained in, biases that exist in all LLMs, and that the ability to interpret human biases against those of LLMs is a vital human-AI collaboration skill. Mr Asensio also stated that by understanding the ‘flow of money’ and data behind these models, students can become more critical consumers and creators of technology.

The panel emphasised that as AI lowers the floor to entry for technical tasks, the ceiling for high-level creative and strategic thinking is raised significantly. The future of technical expertise, specifically coding and digital creation, was debated with a focus on the rise of ‘vibe coding’ and low-code platforms. Dr Frey predicted that within a few years, the demand for traditional entry-level programmers might decrease as AI allows individuals with minimal technical training to build complex applications. However, the panel was quick to clarify that this does not mean technical knowledge is becoming irrelevant; rather, its nature is changing toward architectural oversight and system auditing. Dr McMinn argued that a deep understanding of technological foundations and models developed remains essential to prevent their misuse, even ensuring that they will be used effectively. Mr Asensio noted that in the corporate world, the ability to bridge the gap between technical potential and business value is the most sought-after leadership trait. 

Amidst the shift of skill turns to AI and the contentious space of technical versus human expertise, what should we teach students? There was a consensus among the panellists that banning AI is likely a futile strategy, as students will inevitably interact with these tools in their private lives. Instead, they advocated for a balanced approach to learning with high-tech AI integration and low-tech analogue activities like handwriting and whiteboarding. These traditional methods are seen as essential for maintaining the cognitive pathways required for deep reflection and self-regulated learning. Mr Asensio invoked the classical concept of Sophia (σοφία) or wisdom, the highest form of knowledge according to Greek philosophy, as the union of practical skill and theoretical knowledge as the pathway for the future of education, focusing on wisdom rather than just information. The session concluded with the idea that the ‘AI turn’ should not be feared as a disruptor, but embraced as an opportunity to return to the most fundamental aspects of human intelligence. The final takeaway was that while AI can provide output for the ‘what’ and the ‘how’, it remains the human’s role to judge and answer with the ‘why’ and the ‘should’.

Watch the full panel presentation at ACE2025 in the video below.

Banner image: Nguyen Dang Hoang Nhu, Unsplash