UPI Study › Courses › Introduction to Artificial Intelligence › Free lessons › Lesson 1.1 Free with a free account

What Is Artificial Intelligence? Origins, Ideas, and Growth

Name: 1.1 What Is Artificial Intelligence? Origins, Ideas, and Growth — Introduction to Artificial Intelligence
Uploaded: 2026-10-20T01:49:55+00:00
Duration: 15 min 33 s
Description: Four definitions of AI, the Turing test, and its booms and busts

Introduction to Artificial Intelligence · Chapter 1: Foundations of AI and Core ConceptsSection 1.1 · lesson 1 of 6916 min · free with a free account

Next lesson: Categorizing AI: Types and Distinctions →

Earn 3 college credits for Introduction to Artificial IntelligenceRecommended for credit by NCCRS · Computer Science 373 · self-paced · one payment, lifetime access

Enroll in Introduction to Artificial Intelligence · $250 →Download the syllabus (PDF)

In this lesson

Four definitions of AI, the Turing test, and its booms and busts.

Explain the four ways experts define artificial intelligence using thinking versus acting and humanly versus rationally
Describe how the Turing test works and why passing it does not prove understanding
Trace AI's history from the 1956 Dartmouth workshop through two AI winters to the machine learning revival
Identify the three drivers of AI's growth after 2010 and how they reinforce each other

Free study notes

Get this lesson's notes by email

The key points and 4 practice questions with answers, in your inbox. One email, free.

One email. Unsubscribe in one click.

Transcript

Everything said in the video, section by section (2,092 words). Read along, or search it.

Opening Question

Everyday Tasks Software Now Handles

Ask your phone to find every photo of your dog. Ask a map app for the quickest way home. Ask a chatbot to explain a tricky idea. Each time, software does something we once believed only people could do. So what exactly counts as artificial intelligence? Where did the idea come from? And why does it suddenly seem to be everywhere?

Roadmap

Four Parts From Definitions to Boom

We'll answer in four parts. First, the four ways experts define AI. Then, the test Alan Turing proposed. Then, the story of AI's booms and busts. And finally, the three forces behind today's boom.

Definition

Machines Doing Tasks That Need Human Intelligence

A simple working definition: artificial intelligence is the field that builds machines and software, able to do tasks that normally need human intelligence. Understanding language, recognizing images, making decisions, and solving problems. That sounds clear, but it hides a difficulty. People don't agree on what intelligence is, so they don't agree on what an intelligent machine should be. Experts' definitions fall into four groups, and the groups come from two simple questions.

Two Sorting Questions

Thinking or Acting, Humanly or Rationally

Question one: is the goal about thinking, or about acting? Thinking looks inside the system, at its reasoning. Acting looks at what it does, its behavior. Question two: is success measured against people, or against an ideal? Humanly means copying how people think or behave. Rationally means doing the best possible thing, given what the system knows.

Four Groups

Four Definitions With Examples

Cross the two questions, and you get four groups. Thinking humanly asks: does the program reason the way people do? A model that predicts which words a person will forget from a list is checked against psychology experiments. Thinking rationally asks: does it follow correct, logical steps? All humans are mortal. Socrates is human. So Socrates is mortal. Acting humanly asks: does it behave so that people can't tell it from a person? Picture a chatbot so natural you suspect someone is typing. And acting rationally asks: does it pick the action with the best expected result? A navigation app choosing the fastest route from live traffic is the classic case.

Why Rationality Won

Acting Rationally Drives Modern AI

Most modern AI is built around acting rationally. Here, rational doesn't mean calm or unemotional. It means choosing the action that's expected to give the best result. Three reasons it won. First, it can be measured. Best result can be written as a number, like the shortest travel time, so engineers can tell whether a system is improving. Second, it doesn't require copying the human mind. Engineers didn't build airplanes that flap like birds. They studied lift, and built something that flies well in its own way. Third, it handles uncertainty. Logic alone needs facts that are certain, but the real world is full of guesses. Acting rationally lets a system choose the best action, even when it can't be sure.

Turing Test

The 1950 Imitation Game

In 1950, the British mathematician Alan Turing published a paper called Computing Machinery and Intelligence. It opened with a question: can machines think? Turing felt the question was too vague, because nobody agrees what think means. So he replaced it with a practical one, which he called the imitation game. Today it's known as the Turing test. A human judge sits in a separate room, and types questions to two hidden participants. One is a person, and one is a machine. The judge may ask anything, from what's your favorite food, to what's 17 times 23. Then the judge must decide which one is the machine. If the judge can't reliably tell them apart, the machine passes.

Harder Version

The Total Turing Test

Turing chose a typed chat on purpose. It hides the machine's body, so the judge can only weigh language, knowledge and reasoning. Later researchers proposed a harder version, the total Turing test. The judge also sees video of the participants, and can hand them objects. So the machine must see, and move things, as well as talk.

Strengths and Limits

ELIZA, Searle, and Sounding Intelligent

The test has real strengths. It gave AI a clear, testable goal, instead of an endless argument about one word. One conversation can probe language, knowledge, reasoning and learning at once. And its human-level target pushed early researchers to aim high. But it judges only behavior. A machine might sound convincing without understanding anything, and the result depends on the judge. Skilled or suspicious judges are harder to fool. It also rewards tricks. In 1966, a chatbot called ELIZA simply turned your sentences back into questions, like a therapist, and many people felt it understood them. And in 1980, the philosopher John Searle imagined a person in a room, following a rulebook to answer in Chinese without knowing any Chinese. Fluent answers, he argued, don't prove understanding. So most engineers today don't treat the Turing test as their goal. It remains a famous starting point, and a reminder that sounding intelligent and being intelligent aren't always the same thing.

Early Pioneers

McCulloch, Pitts, Shannon, and the Child Machine

Turing wasn't working alone. In 1943, Warren McCulloch and Walter Pitts proposed a simple mathematical model of a brain cell, and showed that networks of such units could carry out logic. In 1950, Claude Shannon described how a computer could be programmed to play chess, which made games a testing ground for machine thinking. And in the same paper as his test, Turing answered the objection that machines only do what we tell them. Build a simple child machine, he suggested, and let it learn from experience. That sounds very modern.

Quick Check

Classifying a Voice Assistant Test

Quick check. A company tests its voice assistant by playing its replies to listeners, and asking one thing: does this sound like a real person? Which of the four groups is the assistant being judged in? Pause, and decide. Acting humanly. Step one: it's judged on what it does, its replies, so it's acting, not thinking. Step two: the yardstick is people, so it's humanly. Thinking humanly is the tempting choice, but nobody looked at how it reasons. And notice: score the same assistant on how often it gets your request right, and it moves to acting rationally. The group depends on what's being measured.

History Overview

AI Summers and AI Winters

Now the story of how we got here. AI's history doesn't climb in a straight line. It moves in waves. Periods of great excitement and generous funding, sometimes called AI summers, have been followed by disappointment and funding cuts. Those are called AI winters. There have been two big winters so far, and each one has a lesson in it. After each winter came a revival, and the latest one is still going.

Founding Moment

Dartmouth 1956 and Early Optimism

It starts in 1955, when four researchers, John McCarthy, Marvin Minsky, Nathaniel Rochester and Claude Shannon, proposed a summer workshop at Dartmouth College. Their proposal claimed that every feature of intelligence could, in principle, be described so exactly that a machine could imitate it. McCarthy chose the name: artificial intelligence. At the 1956 workshop, Allen Newell and Herbert Simon showed the Logic Theorist, a program that proved mathematical theorems. It's often called the first AI program. Progress then looked fast. Programs solved algebra word problems. Arthur Samuel's checkers program improved by playing games. The Shakey robot at the Stanford Research Institute planned simple moves around a room. Confidence soared. In 1965, Herbert Simon predicted that within twenty years, machines would be able to do any work a person can do.

First AI Winter

Toy Problems Failed to Scale

That prediction didn't come true. Early programs worked on small, tidy toy problems, and struggled as problems grew. The number of possibilities exploded, and computers of the time had little speed or memory. In 1969, the book Perceptrons, by Marvin Minsky and Seymour Papert, showed the limits of the simplest neural networks, and cooled interest in them. In 1973, a UK government report by Sir James Lighthill judged AI's results disappointing, and funders on both sides of the Atlantic cut back. From about 1974 to 1980, research slowed. That was the first AI winter.

Expert Systems

IF-THEN Rules, MYCIN, and XCON

AI came back with a more modest, practical idea: the expert system. It captures the knowledge of human specialists as many IF-THEN rules, so a computer can give expert-style advice in one narrow area. For example: if a customer orders a disk drive, and the order has no controller card, then add a controller card. MYCIN, built at Stanford in the 1970s, suggested treatments for blood infections. Around 1980, a system called XCON began choosing the parts for computer orders at Digital Equipment Corporation. It was reported to save millions of dollars a year. Businesses rushed to build AI teams, and in 1982 Japan launched its Fifth Generation computer project.

Second AI Winter

Brittle Rules and a Collapsed Market

The boom faded, for three reasons. Expert systems were brittle: fine inside their narrow area, but failing on anything their rules didn't cover. They were costly, because people had to write and update thousands of rules by hand, and the systems couldn't learn. And the market for specialized AI computers collapsed, once ordinary desktops were powerful enough. From about 1987, funding dropped again: the second AI winter.

Comparison

What the Two Winters Share

Look at the two winters side by side. Each began after big promises, followed by systems that couldn't deliver outside narrow situations. The first: methods that worked on toy problems failed to scale. The second: brittle, hand-built rules were too costly to keep up. Either way, expectations rose far above what the technology could do, and funding fell when results disappointed.

Machine Learning Revival

From Deep Blue to the Transformer

During the second winter, researchers changed direction. Instead of writing rules by hand, they let programs find patterns in data. That's machine learning. It quietly produced useful tools: speech recognition, spam filters, and product recommendations. Then came headline moments. In 1997, IBM's Deep Blue beat world chess champion Garry Kasparov. In 2011, IBM's Watson won the quiz show Jeopardy. In 2012, a system called AlexNet won a major image-recognition contest by a wide margin. It used deep learning: large neural networks with many layers. In 2016, AlphaGo defeated Lee Sedol, one of the world's best Go players. In 2017, researchers introduced the transformer, the design behind modern language systems. And by 2022, chatbots built on it reached millions of ordinary users.

Three Drivers

Data, Computing Power, and Better Algorithms

Many of the core ideas are decades old. So why did AI grow so fast after 2010? Think of cooking a big meal. You need ingredients, a stove that's powerful enough, and a good recipe. Data is the ingredients. The internet, smartphones, online shopping and cheap sensors created huge amounts of photos, text, voice and clicks. ImageNet, launched around 2009, sorted millions of photos into thousands of categories. Computing power is the stove. Graphics chips, GPUs, made for video games, do thousands of small calculations at once. AlexNet was trained on just two of them. Cloud computing then let anyone rent huge amounts of power by the hour. Better algorithms are the recipe. Training methods kept improving, from backpropagation, popularized in the 1980s, to the transformer. And free, shared software let thousands of people build on each other's work.

Feedback Loop

Each Driver Feeds the Others

And the three feed one another. Cheaper computing makes more data practical. More data makes new algorithms worth inventing. Better algorithms make the most of the hardware. Success then attracts investment, which pays for even more of all three.

Worked Example

ImageNet Error Falls Below Human Level

Here's the result in one field. In the yearly ImageNet challenge, the winning system's top five error was about 26 percent in 2011. Deep learning arrived in 2012, and the error fell to 15 percent, then 12, then under 7. By 2015, the winner made fewer mistakes than the estimated rate for a person, about 5 percent.

Quick Check

Too Few Labeled Insurance Forms

One more check. A team in 2025 wants software that reads handwritten insurance forms. They rent plenty of GPUs in the cloud, and use a modern training method. But they have only 200 labeled forms. Which of the three drivers is holding them back? Pause, and decide. Data, the ingredients. Learning systems improve by studying many examples, and two hundred forms is far too few to find reliable patterns. Computing power is the tempting answer, since big systems need GPUs. But they've rented plenty, and their method is modern. Without enough examples, the stove and the recipe have nothing to cook.

Key Takeaways

Definitions, Turing, Winters, and the Boom

To sum up. AI can be defined four ways, by asking thinking or acting, and humanly or rationally. Most modern AI aims to act rationally. Turing swapped can machines think for the imitation game. It's simple and memorable, but it measures behavior, not understanding. AI was named at Dartmouth in 1956, and went through two winters, each after promises outran what the technology could deliver. And the recent boom came from data, computing power and better algorithms arriving together, and feeding one another.

Reflection

Which AI Claims Deserve Checking

Here's something to think about. Both winters followed promises that technology couldn't keep. Which claims about AI today do you think deserve the most careful checking, and how would you check them?

About Introduction to Artificial Intelligence at UPI Study

Introduction to Artificial Intelligence (Computer Science 373) is a self-paced online course of 69 video lessons in 9 chapters, from intelligent agents and search to game strategy, neural networks, logic and reasoning, machine learning, language, robotics and AI ethics. It carries a NCCRS credit recommendation for 3 semester credits.

Every lesson transcript is free on pages like this one, and the videos are free with a free UPI Study account. Enrolling adds a practice quiz after each lesson, the graded work and final exam that earn the credit, lifetime access, and an official transcript on completion. UPI Study has helped over 130,000 students with college credits. ACE and NCCRS have over 2,100 cooperating colleges that will consider accepting credits; the final decision is with the universities.

Enroll now → Course details Download the syllabus All Computer Science courses Every course with PRO →

Compare: AI vs Current Trends in CS and IT

Why take it here

Why students take Introduction to Artificial Intelligence with UPI Study

How machines search, reason and learnIntroduction to Artificial Intelligence covers how AI systems are built: intelligent agents, search and heuristics, game strategy, logic, neural networks, machine learning, language and perception, and the ethics of AI. It is a common upper-level course in computer science degrees.

Credit at a fraction of the priceThe same 3 credits can cost $750 or more at a US college. Here it is one payment, self-paced, with lifetime access.

Finish your degree soonerTransfer credits reduce the courses you still owe, so you can graduate earlier or lighten a heavy semester.

Strengthen an applicationA completed college-level course on an official transcript shows admissions and scholarship committees you can do the work.

Study on your own scheduleNo start dates, no deadlines. Watch the lessons, practice, and take the final when you are ready.

130,000+students helped with college creditsACE and NCCRS have over 2,100 cooperating colleges that will consider accepting credits; the final decision is with the universities.

Where the credit goes

Universities that work with NCCRS

Introduction to Artificial Intelligence carries a NCCRS credit recommendation. These universities work with NCCRS; some are UPI Study's signed partners. ACE and NCCRS have over 2,100 cooperating colleges that will consider accepting credits; the final decision is with the universities.

Thomas Edison State UniversitySigned partner

University of the PeopleSigned partner

Charter Oak State CollegeSigned partner

Strayer UniversitySigned partner

Columbia Southern UniversitySigned partner

Indiana TechSigned partner

Penn StateWorks with NCCRS

Southern New Hampshire UniversityWorks with NCCRS

SUNYWorks with NCCRS +many more cooperating universitiesFind my college →

Find my college → Not listed? Ask Athena in the corner of this page, or book a 15-minute call.

FAQ

Questions students ask us

About What Is Artificial Intelligence? Origins, Ideas, and Growth

What are the four main definitions of artificial intelligence?

Definitions of AI fall into four groups formed by crossing two questions: is the goal thinking or acting, and is success measured against humans or against an ideal of rationality? That gives thinking humanly, thinking rationally, acting humanly, and acting rationally. Most modern AI follows acting rationally, which means choosing the action expected to produce the best result, such as a navigation app picking the fastest route from live traffic.

What is the Turing test in artificial intelligence?

The Turing test checks whether a machine can pass for a person in typed conversation. Alan Turing proposed it as the imitation game in his 1950 paper Computing Machinery and Intelligence. A judge questions a hidden human and a hidden machine, and if the judge cannot reliably tell them apart, the machine passes. Because it measures only behavior, passing does not prove the machine understands anything.

What caused the AI winters?

Both AI winters followed big promises that the technology could not keep outside narrow situations. The first, roughly 1974 to 1980, came when methods that worked on toy problems failed to scale, with the 1969 book Perceptrons and the 1973 Lighthill report adding to the doubt. The second, from about 1987, came when expert systems proved brittle and costly to maintain and the market for specialized AI computers collapsed.

Why did artificial intelligence grow so quickly after 2010?

Three forces arrived together: data, computing power, and better algorithms. The internet, smartphones, and cheap sensors produced huge datasets such as ImageNet. GPUs, first built for video games, run thousands of small calculations at once, and cloud computing made that power easy to rent. Improved methods, from backpropagation to the 2017 transformer, made the most of both. Each driver strengthens the others, and success attracts more investment.

About the Introduction to Artificial Intelligence course

How many college credits is Introduction to Artificial Intelligence?

Introduction to Artificial Intelligence (Computer Science 373) carries a NCCRS credit recommendation for 3 semester credits. ACE and NCCRS have over 2,100 cooperating colleges that will consider accepting credits; the final decision is with the universities.

How much does Introduction to Artificial Intelligence cost and what is included?

$250, paid once, with lifetime access: all 69 video lessons, a practice quiz after each, the graded work and final exam, and an official transcript when you finish. Students who want several courses choose UPI Study PRO instead.

Which universities accept UPI Study credits?

UPI Study has signed partner agreements with universities such as Thomas Edison State University, University of the People and Charter Oak State College, and works with universities such as Penn State, SNHU and SUNY. ACE and NCCRS have over 2,100 cooperating colleges that will consider accepting credits; the final decision is with the universities. Use "Find my college" on this page to check yours.

How long does Introduction to Artificial Intelligence take?

It is self-paced with no deadlines. The 69 lessons total about 16 hours of video; most students add reading and practice time and finish in a few weeks to a few months, on their own schedule.

How are the exams done?

Quizzes and assignments are online in the course. The final exam is taken online with proctoring; a person on the UPI Study team checks assignments, usually within two days.

How do my credits reach my college?

When the course shows 100% complete, request the official transcript from your dashboard. It is sent to the college you name, and you can send it to more than one. Check your college first with "Find my college" on this page, or ask Athena.

Ready for the full course? Earn 3 credits in Introduction to Artificial Intelligence.

69 video lessons, a quiz after each one, and the graded work that earns the credit. Recommended for credit by NCCRS.

Enroll in Introduction to Artificial Intelligence →Download the syllabus Every course with PRO →

Introduction to Artificial Intelligence
3 credits · NCCRSEnroll →