A human figure stands between stacks of books and a glowing AI circuit network, arms open toward both sides - symbolizing the balance between human knowledge and artificial intelligence in the age of learning autonomy.

When Machines Learn Faster Than We Do: The Future of Human Agency

A new study in the ELT Journal revisits an idea that may soon define what separates human intelligence from artificial systems: agency - the sense of ownership and control over one's actions. As generative AI reshapes education, researchers Jian Tao and Xuesong Gao argue that both learners and teachers must reclaim their role as active decision-makers, not passive processors. Their analysis offers a deeper question for the future of learning itself: what does it mean to act intentionally in a system increasingly designed to act for us?

By Lorans I. Hedgecock October 27, 2025 in Cognitive Science


In the age of generative AI, it has become easy to confuse knowledge with capability. Large language models can now summarize, correct, and even generate essays more fluently than most students. Yet as these systems grow in complexity, a new tension emerges - between automation and agency.

A recent open-access article by Jian Tao and Xuesong Gao in the ELT Journal explores this problem in the context of English language education, but its implications reach far beyond classrooms. Their central idea is simple yet profound: agency is the core of meaningful learning. It is not just what we do, but how consciously we choose to do it.


Agency as Ownership of Mind

The authors define learner agency as "the sense of ownership and control learners have regarding their learning." In practical terms, it's the difference between repeating instructions and deciding how to learn. Teacher agency mirrors this: the ability to shape one's teaching philosophy rather than merely implement a prescribed curriculum.

Agency transforms both teaching and learning into intentional acts. It positions humans not as data processors but as meaning-makers. When this ownership disappears, learning becomes mechanical - the human mind reduced to an input-output loop.

This notion feels especially urgent today. Tao and Gao note that "recent technological developments such as the rise of generative artificial intelligence make it increasingly necessary for educators to articulate and communicate the value of language study." In other words, AI challenges us to remember why we learn at all.


The Hidden Dimensions of Agency

Agency is often mistaken for visible activity: speaking up in class, volunteering, or producing output. But the study reminds us that agency can also appear as silence, reflection, or restraint - the invisible choices that reveal alignment between one's values and actions. A student who chooses not to speak, or a teacher who refuses to enforce punitive rules in favor of empathy, may be exercising a higher form of agency than those who act without awareness.

In this sense, agency is less about motion and more about intention. It marks the difference between movement and meaning - between being animated by the system and animating the system from within.


AI and the Fragility of Choice

Modern educational systems risk eroding this subtle layer of autonomy. Algorithmic tutors, predictive grading systems, and adaptive learning apps claim to "personalize" learning, but often replace agency with optimization. Students are guided toward the statistically most efficient answers, while teachers are evaluated by metrics that privilege conformity.

The question, then, is not whether AI can support learning, but whether it can support freedom in learning. Agency requires friction - the unpredictable space where one's decisions might fail, evolve, or reveal new insight. Remove that friction, and education becomes merely efficient, not transformative.


Reclaiming the Human Space

Tao and Gao propose two key strategies to cultivate agency - both of which apply as much to cognitive systems as to classrooms.

First, they advocate for discursive spaces: environments where teachers (and by extension, all thinkers) can critically reflect on the systems shaping them. These are the intellectual equivalents of neural feedback loops - places where information isn't just received but interpreted, questioned, and restructured.

Second, they highlight the role of multilingual and multimodal reflection - encouraging learners and educators alike to draw on their full cognitive repertoire. This resonates with a deeper truth: agency flourishes when the mind integrates diverse perspectives rather than submitting to a single algorithmic path.


Collective Agency: Beyond the Individual

Interestingly, the authors also describe a form of collective agency, emerging when individuals unite around shared intention. In neuroscience terms, this mirrors distributed cognition - a system where the intelligence of the group exceeds the sum of its parts.

In education, collective agency means teachers collaborating across disciplines, languages, and institutions to challenge limiting paradigms. In society, it might mean reclaiming the right to question the narratives written by algorithms. Agency, at its highest level, is not isolation but co-creation.


The Meaning of Agency in the Machine Age

When viewed through a wider cognitive lens, agency represents more than educational reform. It represents the boundary of consciousness itself - the ability to choose meaning within structure. In the emerging dialogue between human and artificial intelligence, agency is what keeps the human side awake.

The rise of AI does not threaten agency; it tests it. Machines can predict, process, and pattern, but they cannot intend. That remains uniquely human - the spark that turns instruction into insight, repetition into evolution, and data into destiny.


References

Jian Tao, Xuesong (Andy) Gao (2025). Agency. [ELT Journal] https://doi.org/10.1093/elt/ccaf043...

Leave a Comment


How the FDA and EMA Differ on Regulating AI in Drug Development
Nov 9, 2025 Ethics & Governance

How the FDA and EMA Differ on Regulating AI in Drug Development

A new open-access article in The Journal of Law and the Biosciences examines how the US Food and Drug Administration (FDA) and the European Medicines Agency (EMA) are regulating artificial intelligence in drug development. Written by Gabriela Lenarczyk, Timo Minssen, Nicholson Price, and Arti Rai, the study compares the FDA's flexible, dialogue-based oversight with the EMA's structured, risk-tiered model under the EU's new AI Act. The analysis highlights growing divergence across the Atlantic and its implications for innovation, safety, and global harmonization.

AI Summaries May Make Learning Easier - But Shallower, Study Finds
Oct 28, 2025 Creativity & Performance

AI Summaries May Make Learning Easier - But Shallower, Study Finds

A large-scale study published in PNAS Nexus finds that when people learn about a topic through large language model (LLM) summaries - like ChatGPT or Google's AI Overview - they tend to develop shallower knowledge than those who learn through traditional web searches. The research suggests that while LLMs save time, they may reduce the depth of understanding by removing the effort of discovery and synthesis that underpins true learning.

How the Brain Learns: The Hidden Architecture of Reward and Connection
Oct 18, 2025 Cognitive Science

How the Brain Learns: The Hidden Architecture of Reward and Connection

From food to friendship, the human brain is constantly learning what to seek and what to avoid. But new research suggests that reward - the driving force behind learning - is not just an external event or a burst of dopamine. It begins deep inside the body, where every meal, heartbeat, and breath shapes the way we learn, connect, and evolve. A trio of recent papers in Trends in Cognitive Sciences is reshaping how neuroscientists understand learning and motivation - revealing that reward signals arise from internal physiological states, extend through social dynamics, and may hold the key to more natural forms of artificial intelligence.

The Art of Explanation: Three Ways Humans Create Understanding
Oct 25, 2025 Theory & Systems

The Art of Explanation: Three Ways Humans Create Understanding

Every conversation hides a miracle: one mind trying to help another see. According to a new paper in The Philosophical Quarterly, explanation isn't just about giving information - it's about building understanding. Philosophers Grzegorz Gaszczyk and J. P. Grodniewicz propose that explaining is a practice with structure and purpose. Like music, it comes in distinct movements: the minimal, the customized, and the interactive explanation. Each serves a different rhythm of thought - and together, they reveal how humans make meaning visible.

How Users Manage Privacy on Algorithmic Social Media Differs Across Cultures
Nov 3, 2025 Creativity & Performance

How Users Manage Privacy on Algorithmic Social Media Differs Across Cultures

A new open-access study in Journal of Computer-Mediated Communication explores how social media users across the United States and Singapore navigate privacy in an age dominated by AI-driven algorithms. Researchers identified four distinct privacy management profiles - ranging from highly proactive "Privacy-Benefit Maximizers" to disengaged "Privacy Apathy" users - each shaped by confidence, motivation, and cultural context. The findings reveal that privacy behavior is neither uniform nor purely reactive but part of a complex balance between control, convenience, and trust in algorithmic systems.