I must admit, this topic has drifted quite far from my original “AI and Education” series. In the previous post on Ukraine, we touched on words like accountability, transparency, and even the trolley problem. So before we go any further, I thought it would be best to take a step back and ask: What exactly is AI?

This time, I’m writing for adults — especially those who have heard of ChatGPT but aren’t quite sure what “AI” really means. The series will have eight parts: the first six will explore AI itself, and the final two will focus on how to use tools like ChatGPT effectively in daily work and learning.

And yes — rest assured — this “AI and Education” series will indeed end on an education-related theme!

---

🎯 Goal of This Article

To understand the history of AI — how it began, evolved, and reached today’s age of generative AI.

---

Table of Contents

1. What Is AI?
2. The Dartmouth Conference
3. Strong AI vs. Weak AI
4. The First AI Boom: Reasoning and Search
5. The Second AI Boom: The Era of Knowledge
6. The Third AI Boom: Big Data and GPUs
7. Deep Learning
8. The Fourth AI Boom: Generative AI
9. Ethical Challenges and Human Decision-Making

---

What Is AI?

AI — or Artificial Intelligence — refers to “the attempt to replicate human intellectual activities such as understanding, reasoning, and learning using machines.”
The term itself dates back to 1956, nearly seventy years ago.

---

The Dartmouth Conference

In 1956, a group of researchers gathered at Dartmouth College and coined the term Artificial Intelligence. Since then, AI has gone through repeated cycles of “booms” and “winters” — waves of high expectation followed by disappointment when progress slowed. Understanding this history helps us see why AI works now and where its limits still lie.

---

Strong AI vs. Weak AI

AI can be divided into two broad categories:

• Strong AI (General AI): a hypothetical AI that possesses genuine consciousness and can perform any intellectual task a human can — think of Doraemon.
• Weak AI (Narrow AI): specialized systems designed for specific tasks without real consciousness — all the AI tools we currently use, from ChatGPT to image generators, belong here.

---

The First Boom: Reasoning and Search (1950s–60s)

Early AI research focused on solving clearly defined problems like mazes and equations through logical reasoning and search algorithms. However, researchers soon hit the wall of combinatorial explosion — too many possibilities for computers to process. Limited computing power led to the first AI “winter.”

---

The Second Boom: The Knowledge Era (1980s)

The next wave arrived with expert systems, which stored specialists’ knowledge in if–then rules. These systems achieved success in medical diagnosis and industrial maintenance. Yet, capturing human expertise (especially tacit knowledge) proved costly and difficult. The so-called “knowledge acquisition bottleneck” caused another slowdown.

---

The Third Boom: Big Data and GPUs (2000s–)

The landscape changed dramatically in the 2000s with three breakthroughs:

• The rise of big data from the internet
• GPUs capable of massive parallel computation
• New machine learning algorithms

This allowed computers to learn patterns directly from data rather than relying on hand-crafted rules.

---

Deep Learning

Among machine learning methods, deep learning — multi-layered neural networks inspired by the human brain — revolutionized image and speech recognition. For the first time, machines could automatically extract features from raw data, achieving superhuman accuracy in certain tasks.

---

The Fourth Boom: Generative AI (2020s–)

We’re now in the fourth wave — the age of generative AI. Tools like ChatGPT, Gemini, and Canva’s AI can produce text, images, and even music. Instead of just analyzing data, AI now creates new content.

This shift marks AI’s transformation from a lab demo to a practical assistant that boosts productivity and quality in everyday work.

---

ChatGPT and the Start of a New Era

Many mark November 2022, the launch of ChatGPT, as the dawn of the generative AI boom. ChatGPT made it possible for anyone to interact with a large language model (LLM) — the “brain” of modern AI — through natural conversation.

---

Challenges That Remain

AI still struggles with deep, human-like understanding.
Some classic unresolved issues include:

• The Frame Problem: determining what’s relevant or irrelevant in a situation.
• The Symbol Grounding Problem: connecting symbols (like words) to real-world meaning.
• Moravec’s Paradox: AI can play chess but struggles with basic sensorimotor tasks that toddlers perform easily.

These challenges remind us that AI works best when humans define clear goals, curate data, and make final decisions.

---

Historical Use Cases by Sector

Public Administration

Rule-based AI (2nd boom) still helps check application formats or legal compliance. Generative AI (4th boom) now drafts public announcements and FAQs, improving consistency and reducing waiting times.

Healthcare

Machine learning enhanced diagnostic imaging (3rd boom), while generative AI assists with medical notes, referrals, and summaries — allowing doctors to focus on communication and ethical decision-making.

Manufacturing

AI detects anomalies in sensor data and drafts inspection reports. Combined with MLOps (AI model operations), this enables predictive maintenance and smoother production lines.

Education

Learning analytics (machine learning) identify where students struggle. Generative AI restructures lecture notes into summaries, examples, and quizzes — freeing teachers to spend more time on personalized interaction.

---

AI Ethics: The Trolley Problem

A classic thought experiment in ethics — the Trolley Problem — asks whether you would switch the tracks of a runaway trolley to save five people at the cost of one.
As AI systems gain autonomy (e.g., self-driving cars or medical triage), who decides such moral priorities?

The key lies in creating transparent governance:

Define principles → Implement them in AI → Record decisions → Continuously revise through social consensus.

---

Humans Must Make the Final Call

AI’s seventy-year journey teaches us not to fear or idolize it.
When we treat AI as “a first draft and a planner” — not the final decision-maker — it becomes a powerful ally.

We now stand in a unique era: technology, data, and computational power are finally aligned.
The challenge ahead is to apply the right tool, in the right context, with the right governance.

---

Thank you for reading. If you enjoyed this article, please give it a “like,” leave a comment — I’d love to hear your thoughts — and follow for future posts.

With gratitude,
Warm regards,
Tobira AI