Section 3.1 What is Artificial Intelligence?

Artificial intelligence is one of the most influential technologies in modern life, yet it is often misunderstood as either magical or purely futuristic. In this section, we define what AI actually is, separate narrow AI from general AI, and trace the key historical milestones that shaped today’s systems.

Subsection 3.1.1 Defining AI

We will begin our investigation of this field by addressing a fundamental question: What does it mean for a machine to be "intelligent"? Steve Wozniak, co-founder of Apple computers, reminded college graduates in May 2026 that human beings possess ’Actual Intelligence’.

If you look at early computing devices, they were essentially high-speed calculators. They could follow rigid mathematical instructions perfectly, but they lacked the ability to adapt. When we talk about Artificial Intelligence (AI), we are describing a shift away from pure calculation toward systems that can perceive their environment, learn from experience, and take independent actions to achieve a specific goal. These systems are created by human beings to Augment Intelligence.

Computer scientists typically divide the pursuit of AI into two main philosophies:

Thinking and Acting Humanly: Building systems that mimic human thought processes, cognitive behaviors, and biological structures (like the neural networks in our brains).
Thinking and Acting Rationally: Building systems that rely on formal logic, probability theory, and mathematics to make the best possible objective decision, regardless of whether a human would solve it the same way.

Subsection 3.1.2 The AI Spectrum

When analyzing AI systems in the modern world, it is crucial to understand that not all "intelligence" is scaled equally. We categorize AI capabilities into two distinct levels on a spectrum:

Artificial Narrow AI (ANI): Also frequently called Weak AI. These are software systems designed, trained, and optimized to handle a single, highly specific task. A narrow AI might be incredibly dominant at its given objective, such as a chess algorithm that can defeat a grandmaster, or a recommendation loop that predicts which video you will click on next. However, that chess-playing algorithm cannot translate a sentence, and the recommendation system cannot drive a car. It is entirely bounded by its narrow domain.
Artificial General AI (AGI): Also referred to as Strong AI. This describes a theoretical form of intelligence where a machine possesses a human-level adaptability of mind. An AGI system would not need to be reprogrammed for a new task; it could learn to paint a picture, write code, or reason through an unfamiliar ethical dilemma using general cognitive flexibility. True AGI does not yet exist and remains a theoretical milestone for researchers.

Subsection 3.1.3 A Brief History

How did we arrive at the massive, data-driven AI systems we use today? The field did not appear overnight; it is the result of decades of shifting concepts and computational evolution.

The philosophical foundation was laid in 1950 by British mathematician Alan Turing. He proposed a famous thought experiment known as the Turing Test. Pretend that a human judge is typing messages into a computer terminal, conversing with two hidden entities: a real human and a machine. If the judge cannot reliably tell which entity is the computer after a free-flowing conversation, the machine has successfully demonstrated intelligent behavior. A few years later, in the summer of 1956, a group of scientists gathered at the Dartmouth Workshop. It was at this conference that the term "Artificial Intelligence" was officially coined, launching the field as an independent academic discipline. Early pioneers were deeply optimistic, believing that human-level intelligence could be recreated in a generation using basic symbolic logic.

However, early computers simply did not have the memory or speed to handle the complexities of the real world. This led to decades of stagnant funding and skepticism, historically referred to as "AI Winters."

The modern boom we are experiencing today only became possible due to a perfect convergence: the sudden availability of massive, global datasets to learn from, alongside the development of high-performance parallel computing hardware (GPUs). Instead of trying to hand-code every single rule of logic into a computer, scientists shifted to letting computers parse this massive sea of data to learn the rules for themselves.

Check your understanding

Checkpoint 3.1.2.

In 1997, IBM’s Deep Blue made history by defeating the reigning world chess champion, Garry Kasparov, in a six-game match under standard tournament time controls. Deep Blue was capable of exploring up to 200 million possible chess positions per second to choose its best move, but it could not read an article, converse with a human, or solve a simple tic-tac-toe game.

Deep Blue is an example of which level on the AI spectrum?

Artificial General AI (AGI), because defeating a human world champion in a highly complex intellectual game requires human-level cognitive flexibility.
Incorrect. Achieving mastery or defeating humans in a specific game does not make an AI general. An AGI requires an adaptability of mind to learn entirely unrelated tasks on its own.
Artificial Narrow AI (ANI), because despite its dominance at chess, its capabilities were entirely bounded to a single, highly specific mathematical domain.
Correct! Deep Blue is a classic example of Artificial Narrow AI (ANI). No matter how superior an AI is at its chosen task, if it cannot adapt its intelligence to function in a completely different domain without being entirely rewritten, it remains narrow.
Actual Intelligence, because its decision-making process mimicked the organic biological intuition of human chess grandmasters.
Incorrect. As Steve Wozniak noted, machines lack the organic "Actual Intelligence" of humans. Deep Blue relied on brute-force search algorithms and hardcoded evaluation logic, not human intuition.
An AI Winter, because its reliance on massive computation rather than data-driven learning sets caused stagnation in computer science research.
Incorrect. Deep Blue’s victory was a highly publicized success milestone that generated immense interest in computing, quite the opposite of an "AI Winter."

Checkpoint 3.1.3.

In science fiction, JARVIS from the Marvel Cinematic Universe acts as an intelligent assistant to Tony Stark. JARVIS independently manages complex automated factories, hacks secure global networks, invents entirely new molecular elements, cracks jokes, holds philosophical debates, and autonomously adapts to completely unfamiliar combat threats without needing structural software rewrites.

If JARVIS were to exist in the real world today, how would computer scientists categorize its intelligence?

Artificial Narrow AI (ANI), because it is ultimately just a software co-pilot meant to augment one human being.
Incorrect. JARVIS operates across a limitless number of distinct tasks, meaning its domain is general rather than narrow, regardless of who it is assisting.
Symbolic Logic AI, because its ability to speak English perfectly proves it relies entirely on rigid, pre-programmed grammatical rules.
Incorrect. Symbolic logic alone was historically unable to handle the chaotic complexities of language or general problem-solving, which is what led to the original AI winters.
Artificial General AI (AGI), because it demonstrates a human-level adaptability of mind, emotional awareness, and the general cognitive flexibility to learn and solve any intellectual task.
Correct! Fictional AIs like JARVIS, HAL 9000, or Commander Data represent Artificial General AI (AGI) because they exhibit general cognitive flexibility, contextual reasoning, and independent adaptability across entirely unrelated fields.
Weak AI, because a machine operating in a fictional universe cannot meet Alan Turing’s real-world definition of passing the Turing Test.
Incorrect. Fictional portrayals are used as conceptual benchmarks for Strong AI or AGI precisely because their behavior demonstrates full human-level capability and conceptual self-awareness.

You have attempted of activities on this page.

Prev Top Next