In my coursework and research… more often in recent times, I watched a familiar scene unfold, one that reminded me of Mr. Miyagi’s (Pat Morita’s character in 80’s film, Karate Kid) wisdom about learning through doing. A student, faced with a complex problem, reached for ChatGPT rather than struggling through the fundamentals. It’s a pattern I’ve observed intensifying over twenty years of teaching with technology: the shift from technology as a tool to technology as a substitute for essential learning processes. While AI and automation offer unprecedented efficiency in education, my research and classroom experience suggest we’re approaching what I call the “automation abyss” – a fundamental transformation in how humans learn and develop skills. The traditional digital divide of access to technology is evolving into something more subtle: a divide between those who maintain agency over their learning and those who become dependent on AI-mediated experiences. This shift, accelerated by powerful language models and automated learning tools, raises crucial questions about the future of human learning and development.
In one of my first blog posts on this website in 2019, I started to put into words just how automation was going to start molding how we teach and learn. Some of the concepts and concerns here have become much more pronounced in the last 6-7 years.
The Digital Divide becomes the Automation Abyss
Drawing from my extensive work in digital citizenship and educational technology, I’ve observed a profound shift in how we think about the digital divide. While we once focused primarily on access to technology and digital literacy skills, we now face a more complex challenge that I call the “automation abyss” – a new kind of digital divide that separates human-centered and AI-mediated interactions.
As I outlined in my recent research on the Budding Botanist Paradox (Hawkinson, 2022), the more we automate learning processes through AI and immersive technology, the more dependent learners become on these automated systems. This creates a dangerous cycle where “the more automated the learning process becomes, enhanced and augmented with immersive technology, the more learners could be dependent on these automated systems for basic learning”.
The Budding Botanist Paradox: Automating Human Inquiry with Immersive Technology. (2022). International Conference on Computers in Education. https://library.apsce.net/index.php/ICCE/article/view/4550
This evolution mirrors what I’m seeing in my Digital Citizenship and Engagement course at the University of Foreign Studies, as course that will soon be adopted by other universities in other countries. Students increasingly rely on AI tools not just to complete tasks, but as intermediaries for fundamental learning and human interaction. The traditional digital divide of “haves” and “have-nots” is being replaced by a more insidious division – one between those who maintain agency over their learning and those who become dependent on AI-mediated experiences.
The stakes couldn’t be higher. As my research suggests, this isn’t just about educational outcomes. It’s about who controls the fundamental infrastructure of human interaction, learning, and commerce. We’re seeing this play out dramatically with developments like DeepSeek, which has “sparked market turmoil” and raised crucial questions about the control of AI infrastructure. What makes this particularly concerning is how AI content is beginning to saturate our digital spaces. In my classroom observations, I’ve noted that “AI giving similar responses to student prompts get pulled up in class again and again, without much time to ‘sit’ in the issue or struggle in the unknown”. This homogenization of thought and learning represents a fundamental shift from AI as a tool to AI as an intermediary controlling our thought processes.
The automation abyss presents a critical challenge for digital citizenship. As I teach my students, we must maintain human agency in an increasingly automated world while ensuring that technology enhances rather than replaces authentic human interaction and learning. This balance will determine not just the future of education, but the very nature of human society in the digital age.
Beyond the Space Race: Understanding the AI Land Rush in the Learning Space
Drawing from my research and experience in educational technology, I see striking parallels between today’s AI race and the Space Race of the 20th century, but with a crucial difference. While the Space Race was fundamentally about technological supremacy, today’s AI competition centers on controlling the infrastructure of human learning and interaction. The recent emergence of DeepSeek provides a perfect case study. Just as the Soviet Union’s Sputnik moment catalyzed American technological development, DeepSeek’s success has been described as “AI’s Sputnik moment”. However, what’s particularly telling is that DeepSeek achieved comparable performance to leading AI models at “a tenth of the computing power” and significantly lower cost. This isn’t just about technological achievement – it’s about who controls the gateways of automated interaction.
In my work studying the Budding Botanist Paradox, I’ve observed how this mirrors historical patterns of infrastructure control. Just as railroad companies in the 19th century didn’t just sell transportation but controlled the vital arteries of commerce, today’s AI companies aren’t just selling models – they’re establishing themselves as the essential intermediaries of human interaction and learning.
This automation gold rush has three distinct players:
- The miners: Early AI adopters and startups rushing to stake their claims
- The pickaxe sellers: Companies like DeepSeek and OpenAI selling AI models, fueled by chip makers like NVIDIA
- The railroad barons: Big Tech firms positioning themselves to control the automation infrastructure
The consequences of this race are already visible in my Digital Citizenship classroom and others. As I’ve documented, students are becoming increasingly dependent on AI tools, not just for completing tasks but for fundamental learning processes. This mirrors what I call the “paradox of automation dependency” – similar to what we saw with Air France Flight 447, where over-reliance on automation led to tragic consequences when human skills were suddenly needed.
Field experiments in education are starting better confirm my work to show a concerning pattern: while AI tools provide impressive short-term gains, they can often lead to long-term skill erosion. Just as pilots can lose their edge when overly dependent on autopilot, students risk losing crucial critical thinking and problem-solving abilities when too reliant on AI intermediaries. This is why I advocate for a more nuanced approach to AI adoption in education. As I tell my students, it’s not about rejecting automation but understanding its proper role. The real challenge isn’t winning the AI race – it’s ensuring that in our rush to automate, we don’t lose the essential human elements that make learning meaningful and lasting.
As a learning futurist focused on the intersection of technology and education, I believe the recent Penn study on AI’s impact on math learning (See Below) provides crucial validation of what I’ve observed in my Digital Citizenship and Engagement course and other related research. The research demonstrates that while AI tools can dramatically improve immediate performance (by 48-127%), they can actually harm learning outcomes when students lose access to these tools, with performance dropping by 17% compared to students who never used AI assistance.
This mirrors what I’ve termed the “Budding Botanist Paradox“, where increasing automation of learning processes leads to greater dependency on AI systems, especially as we move to use technologies like augmented and virtual reality. This is particularly concerning as we see major developments like DeepSeek pushing us toward more AI-mediated learning environments. In my classroom, I’ve observed how “AI giving similar responses to student prompts get pulled up in class again and again, without much time to ‘sit’ in the issue or struggle in the unknown”. This homogenization of learning experiences threatens the development of critical thinking skills. The challenge ahead lies in designing learning environments that harness AI’s benefits while preserving essential human learning processes.
Navigating the Automation Abyss: Digital Citizenship as Framework
What I am observing in the various institutions I work with, research or coursework related, Is an emerging pattern I call the “automation loop” in education. Teachers are using AI to refine coursework, students are using AI to complete that coursework, and educational technology companies are selling AI products to detect student AI work and then give AI generated feedback. This creates a concerning scenario where authentic human learning interactions become increasingly less common. Through my research on the Budding Botanist Paradox, I’ve found that safeguards are essential but must be implemented thoughtfully. For instance, the Penn study demonstrated that simply adding guardrails to AI tutoring systems (like requiring step-by-step problem solving) can eliminate the negative learning effects while maintaining performance benefits . In my coursework, I’ve developed a framework that emphasizes three key components:

- Required Human Interaction: Designing learning activities that necessitate genuine human-to-human discussion and debate. As I tell my students, “your ability to critically think and exchange your opinions about with others about these topics is more important than being able to write an essay about them”.
- Critical Thinking About Automation: Teaching students to understand when and why to use AI tools, not just how. This aligns with my research showing that “the more automated the learning process becomes… the more learners could be dependent on these automated systems for basic learning”. One example of this trying to build in some struggle into the learning, the essence of game based learning mixed with the hero’s journey, understanding that its not about the destination but the process of asking questions and failing along the way.
- Self-Awareness in Technology Use: Helping students develop metacognitive skills to recognize when technology is enhancing versus replacing their learning. This becomes particularly crucial as we see AI systems like DeepSeek becoming increasingly sophisticated and accessible. One example of this is a module on digital footprints, where I ask students to research and analyze what data has been collected about them online on various outlets.
For educational institutions and policymakers, I recommend implementing what I call “human checkpoints” – crucial moments in the learning process that require authentic human engagement. These might include peer-to-peer teaching, open-ended discussions, or creative problem-solving sessions that can’t be easily automated. Educational technology companies need to shift their focus from pure automation to what I call “augmented pedagogy” – tools that enhance rather than replace human teaching and learning. This means designing AI systems that prompt deeper engagement rather than providing quick answers, similar to the safeguards demonstrated in the Penn study’s “GPT Tutor” system.
Final Thoughts
Education is at a turning point. Artificial intelligence is becoming a central force in classrooms, offering tools that automate tutoring, essay writing, and even grading. This shift raises an urgent question. Are we improving learning or removing the very struggle that makes it meaningful? Efficiency is not the same as understanding. Struggle, uncertainty, and effort build deep learning. AI-generated assignments remove the need for original thought. AI-powered grading tools reduce the role of human mentorship. Students may learn to use AI, but will they learn to think critically without it? This is the Automation Abyss. AI makes learning easier but weakens curiosity, creativity, and resilience. Without grappling with complex ideas or struggling through mistakes, students risk losing the ability to problem-solve and think independently.
Education must not fall into this trap. AI should not replace the learning process. It should challenge students, encourage exploration, and enhance critical thinking. The future of education depends on keeping humans in the loop, not cutting them out of it.
About the Author
Eric Hawkinson
Learning Futurist
erichawkinson.com
Eric Hawkinson is a Learning Futurist at Kyoto University of Foreign Studies, where he focuses on the integration of technology into education. Specializing in the creation of immersive learning environments, Eric employs augmented and virtual reality to enhance learning outcomes. He is an advocate for digital literacy and privacy, promoting open access to information and ethical technology practices. Outside his academic role, Eric is engaged in public outreach and professional development. He has established immersive learning labs, designed online courses, and advised on technology strategies across various sectors. His professional designations include Adobe Education Leader, Google for Education Certified Innovator, and Microsoft Innovative Expert. Eric’s notable projects, such as AR experiences for TEDxKyoto and WebVR for Model United Nations, reflect his commitment to using advanced technologies for global education and collaboration. Eric is dedicated to exploring the challenges and opportunities presented by emerging technologies, contributing significantly to the evolution of educational practices.
Roles
Professor – Kyoto University of Foreign Studies
Research Coordinator – MAVR Research Group
Founder – Together Learning
Developer – Reality Labo
Community Leader – Team Teachers
Co-Chair – World Immersive Learning Labs
Read more:
- Is Education Falling into the Automation Abyss? Why Struggle and Human Connection is Essential for Learning in an AI-Driven World
- TikTok Refugees – A Defining Moment in Social Media Migration
- Year-End Academic and Professional Achievements Report (2024)
- Wax On, AI Off: Learning in the Age of Generative AI
- Partnering with the University of the Philippines: Pioneering Micro-Learning with FASTLearn