Forget the AI Hype: Two Lessons on What Drives Technology Adoption

Introduction

When a new technology like Artificial Intelligence or rather the Large Language Models like ChatGPT, Claude or Gemini enters our work and lives, our attention is often misdirected. We see a familiar pattern: on one side, there is hype and optimism about solving all our problems; on the other, there is anxiety about control and obsolescence. A third group quickly emerges: self-styled experts who profit from this confusion by selling courses on the technical specifications of the new tools, without telling people why and how they should use the new technology.

This is happening now with AI, just as it did with the personal computer. However, the history of technology adoption teaches a clear lesson. The primary obstacles to implementation are not technical; they are psychological and institutional. Success does not come from understanding the machine itself, but from creating an environment where it becomes genuinely useful. Below I describe the two main barriers to technology adoption, and neither is technical.

Educator Liability Under the EU AI Act: Risks and Requirements for AI-Assisted Assessment

 The EU AI Act transforms educators into regulated operators of high-risk AI systems when using artificial intelligence for student evaluation, creating unprecedented personal liability exposures. Under Article 26(2) of Regulation (EU) 2024/1689, teachers and lecturers deploying AI assessment tools assume direct compliance responsibilities that carry financial penalties up to €15 million or 3% of institutional turnover for violations[8][15]. This legal framework repositions classroom AI use from pedagogical choice to regulated activity with consequences extending beyond institutional liability to individual accountability.

The High-Risk Designation of Educational AI Systems

Regulatory Classification

Annex III(3)(b-d) explicitly categorizes AI systems used for:

1. Learning outcome evaluation (assessment)
2. Academic level assignment (student selection and streaming)
3. Exam behavior monitoring (also called proctoring)

as high-risk applications requiring strict compliance[9][12]. The Act presumes these systems impact fundamental educational rights under Article 14 of the EU Charter, triggering employer obligations regardless of system complexity[6][13].

Rewiring Education: AI, Ethics, and the Future of Learning

 

Introduction

The intersection of artificial intelligence (AI), education, and ethics is no longer theoretical. It is happening in real-time—in classrooms, curriculum design sessions, and national policy debates. As AI systems become increasingly embedded in our teaching and learning environments, we must not only adapt but rethink the entire educational framework.

If we treat AI as just another tool, we miss its transformative potential. AI is reshaping how we learn, how we teach, how we assess, and ultimately, how we define education itself. This post explores three key areas where AI is already forcing us to rethink education: pedagogy and system design, the transformation of the teacher’s role, and the ethical scaffolding required to deploy AI safely and effectively.

---

I. From Tools to Systems: Redesigning Learning with AI

In her interview, Sinead Bovell makes a clear point: AI is not merely a classroom supplement—it is reshaping the entire learning ecosystem. This aligns closely with my own experience designing and piloting AI-powered business simulations in high school economics and business management courses using Flintk12.com (AI4TL, 2024).

In my view, 3 main issues needs to address for successfully merging AI into educational frameworks: 

• safe adoption suited for learners, especially minors,
• effective and transparent use,
• clear guidelines, ethical and critical implementation.

This calls for immediate updates to curricula and assessment, and a comprehensive redesign of education systems. 

This emphasizes that AI should enhance, rather than hinder, educational outcomes. Real-time feedback and self-paced learning were identified as beneficial aspects of AI, while the current challenges involve preventing misuse such as cheating and ensuring that deep learning prevails over traditional rote memorization.

A Simulation-Based Approach

Instead of relying on traditional explanatory lectures or group activities, I introduced adaptive simulations using Large Language Models (LLMs) to facilitate decision-making exercises. It was like role playing but with an AI coach providing immediate and personalized feedback. Linked to the unit's learning objectives, Students assumed executive roles in simulated companies, made strategic decisions, and received immediate, personalized feedback from AI. The results were clear: students were more engaged, took ownership of their learning, and developed skills in a context that mirrored real-world complexity.

Bridging the AI Gap in Education through Prompt Design: An Individual Approach

Summary:

• 🏛️ Schools and universities are historically slow to change, creating a gap.
• 🧑‍🎓 Individuals are adapting to AI faster than institutions.
• 🔑 Prompt design is the key skill, not technical AI knowledge.
• 🌟 Strategic AI use can transform both learning and teaching

With Bing Image Creator

Here is my presentation with sample prompt of AI for educators. 

Bridging the AI Gap in Education: An Individual Approach

Schools and universities, despite their enduring legacy, are often slow to adapt to rapid technological change (see my blog post A Stagnant Sea). The advent of readily available large language models (LLMs) in late 2023 has created a significant gap between institutional AI strategies (or lack thereof) and the potential for individual adoption. This post argues that educators and students can, and should, take the lead. The rewards are enormous, ever imagined your life if you could do a week's work in one day?

A Trojan Horse or a Tool for Transformation?

Background: The AI Tsunami

It's undeniable: Artificial Intelligence (AI) tools like ChatGPT have arrived, and they're not going away. Students are already using them. The question is no longer if AI will impact education, but how. 

Many are rushing to "AI detection" software. However, as the recent MIT Sloan article, "AI Detectors Don't Work. Here's What to Do Instead," clearly states, these tools are unreliable, prone to error, and can lead to false accusations of academic dishonesty. OpenAI, the creator of ChatGPT, even shut down its own detector due to poor accuracy. This begs the question: what should middle and high school teachers do?

Created with Bing Image Creator

Apple vs. Windows: My Tech Tango (and Why the iPad Stole My Heart)

Introduction

Let's be honest, sometimes keeping up with the latest technology feels like...well, like getting a root canal. You know it's probably necessary, maybe even beneficial in the long run, but the immediate prospect fills you with a vague sense of dread. There are new acronyms, constantly shifting interfaces, and the nagging feeling that you're one software update away from total technological obsolescence. And for many years, this feeling was amplified by the seemingly endless debate: Apple or Windows? Which ecosystem would reign supreme in my digital life?

With Bing Image Creator

 

Apple vs. Windows

For years, I was firmly in the Windows camp. I grew up with it. I knew its quirks, its strengths, and its (many) weaknesses. I considered myself tech-savvy, an early adopter in many respects, always eager to explore new software and hardware – within the Windows ecosystem. Apple products, on the other hand, seemed…different. Sleek, yes, but also restrictive, expensive, and somehow…too polished. 

DeepSeek AI: How China has Caught Up

Original Title: How China’s New AI Model DeepSeek Is Threatening U.S. Dominance. 

Here is a summary of the Youtube video: Consumer News and Business Channel CNBC. (2025, January 24). How China’s New AI Model DeepSeek Is Threatening U.S. Dominance. YouTube. https://www.youtube.com/watch?v=WEBiebbeNCA&t=336s

A little-known AI lab out of China has ignited panic throughout Silicon Valley after releasing AI models that can outperform America’s best despite being built more cheaply and with less-powerful chips. DeepSeek, as the lab is called, unveiled a free, open-source large-language model in late December that it says took only two months and less than $6 million to build. It does not pass the Tienanmen test (just ask it), but apart from this no so minor detail, it is a reasonably reliable tool.

The new developments have raised alarms on whether America’s global lead in artificial intelligence is shrinking and called into question big tech’s massive spend on building AI models and data centers. In a set of third-party benchmark tests, DeepSeek’s model outperformed Meta’s Llama 3.1, OpenAI’s GPT-4o and Anthropic’s Claude Sonnet 3.5 in accuracy ranging from complex problem-solving to math and coding. 

CNBC’s Deirdre Bosa has the story. This video also includes Bosa’s full interview with Perplexity CEO Aravind Srinivas. Quote: "Look, in general, it's known that once open-source is caught up or improved over closed source software, all developers migrate to that." [00:25:12]

Summary

China's Deepseek AI model has rapidly surpassed U.S. tech giants in efficiency and cost, challenging American dominance in AI development.

Highlights

• 🚀 Technological Leap: Deepseek's AI model outperformed U.S. counterparts in just two months with minimal investment.
• 💰 Cost Efficiency: Deepseek spent only $5.6 million, while competitors like OpenAI and Google invested billions.
• 📊 Performance: Deepseek's model excelled in various benchmarks, outperforming established models like GPT-4 and Llama.