Gains:
- Can distinguish between the concepts of artificial intelligence, machine learning and big language model
- Explain how these technologies differ from traditional software
- Can recognize AI examples encountered in daily business life
In recent years, the word "artificial intelligence" is everywhere: in the news, in advertisements, in business meetings, even in the camera of the phone you use. But what exactly does it mean? This unit explains the subject from the basics, without drowning in jargon, and by connecting it to your business life. The aim is not to make you an engineer; To make you a conscious user who understands what the tools you use are and why they do what they do. Because the better you know a vehicle, the safer and more efficient you will use it.
You often hear three concepts intertwined: artificial intelligence, machine learning, and the big language model. These are not the same thing; They are increasingly narrowing rings, one inside the other. Now let's explain these one by one with daily examples.
Three Concepts, Three Interlocking Rings
Artificial Intelligence (AI): It is the general name of a broad field that enables computers to perform tasks that normally require human intelligence (understanding, decision-making, recognition, using language). A chess program, a translation application, and a voice assistant are also under this umbrella.
Machine learning (ML): It is the most common sub-branch of artificial intelligence. Here, we do not write the rules into the computer one by one; We show him lots of samples and he picks out the pattern in the samples himself. Hence the name “machine learning”: the machine learns from examples.
Large Language Model (LLM): It is a special type of machine learning that specializes in text. It is trained with billions of sentences; It is used in tasks such as understanding the language, producing, summarizing and translating. This is the technology behind tools like ChatGPT, Gemini, Claude. The tools in this module, which we call "generative AI" (text/image generating AI), are mostly LLM based.
Hint: Remember this: AI is the biggest box, machine learning is the box inside it, and the large language model is the innermost box. “Every LLM is a product of machine learning and every machine learning is an AI, but the converse is not always true.”
Difference with Traditional Software: Who Writes the Rules?
This is the most important difference to understand. In traditional software, a human (programmer) writes the rules one by one: "If the order amount exceeds 500 TL, the shipping will be free." The computer follows this rule to the letter, no more, no less.
In machine learning, humans do not write the rules; The model discovers the rule itself from the examples shown to it. For example, you show thousands of "spam" and "not spam" emails; The model itself learns which words and phrases indicate spam. Nobody writes one by one saying "if there is this word, it is spam".
feature
Traditional Software
Machine Learning / AI
Who makes the rules?
Man writes one by one
The model infers from examples
Is the output always the same?
Same input → same output
Same input → similar but variable output
When is it strong?
Work with clear, definitive rules
Works with lots of patterns and difficult rules
reliability
predictable, steady
Statistical, requires verification
example
payroll account
Improve email tone
This difference is the basis of the "hallucination" and "verification" topics you will learn later: Because the model derives the rule itself, it is sometimes wrong, so its output cannot be blindly trusted.
You Already See It in Your Daily Business Life
Don't think AI is something new; You've probably been using it for years:
- Your email box's spam filter (a model that has learned which emails are trash).
- Your phone's keyboard prediction (a small language model that suggests the next word).
- Your bank's fraudulent transaction alert (model that detects unusual spending).
- Your spreadsheet program's autofill and suggestion features.
- Recommendation systems in video and music applications.
What's new is that this technology can now speak with natural language. So you write as if you were talking to an expert, and he understands and answers. This is where the real opportunity lies in the workplace.
Three Mini Cases
Case 1 — Ayşe, office manager. He was manually categorizing around 40 supplier emails each week; It took an average of 90 minutes. It started having the company's AI assistant label emails as "urgent/normal/info." The time was reduced to 20 minutes; He devoted ~70 minutes saved to contract control. The AI learned the rule itself; Ayşe did not write filters one by one.
Case 2 — Mehmet, field technician. The customer was writing fault records in free text (“device gets hot, shuts down intermittently”). His manager could not read and classify these 500 records manually. A language model summarized the records as "hardware/software/user error" and extracted the most frequent complaint. The model sketched in minutes what would take nearly 2 days for a human; the team only confirmed.
Case 3 — Elif, accounting. He had AI summarize a new legislative text and took a 1-page list of articles from the 12-page text. But he himself confirmed a critical percentage from the official source; because it was a matter that required numerical precision. Thus, it worked both quickly and safely.
Weak Prompt / Strong Prompt
Knowing the tool also changes how you ask it. Two different questions from the same assistant:
Weak prompt:What is artificial intelligence?
Result: A long, encyclopedic description that is not specific to you.
Strong prompt: I am a non-technical office worker. Explain the difference between artificial intelligence, machine learning and big language model with 3 short bullet points and an example from daily life for each. Use plain Turkish.
Result: A usable answer appropriate to your role, level and purpose. Difference: you gave context, audience and format.
Copiable Templates
Help me learn a technology.Topic: [e.g. machine learning]Me: [my role / my knowledge level]I want: A simple explanation in 5 items + an example from business life for each item + a 1-sentence "keep it in mind" summary at the end.
Show me where I encounter the following concept in daily life:Concept: [e.g. recommendation system]Give me 5 realistic examples and explain in one sentence exactly what the AI does in each one.
Compare these two concepts and put them in a table:Concept A: [traditional software]Concept B: [machine learning]Columns: who sets the rule, is the output constant, when is it strong, an example.
Simplify the following sentence as if explaining it to a non-technical colleague. If there is jargon, add a short explanation in parentheses. Sentence: [paste here]
Common Mistakes
Common mistakes
- Thinking "AI = robot / consciousness". Today's vehicles do not think or feel; recognizes patterns and generates possible sequels. Forgetting this leads to exaggerated expectations and disappointment.
- Treating every AI output as definitive truth. The model works statistically; It is risky to use numbers, dates and sources that require precision without verifying them.
- Saying "This job is far from me" and never trying. AI saves the most time on tedious and repetitive tasks of non-technical employees. Those who don't try will miss the opportunity.
- Asking one-word questions without context. If you don't tell the tool who you are and what you want, you'll get generic, useless answers.
Caution: AI "learning" is not like human learning. The model "learns" not by understanding the world but by capturing statistical patterns in the data. So he can speak very fluently but still get a basic fact wrong. Fluency is no guarantee of accuracy.
In summary
- Artificial intelligence is the broadest umbrella; machine learning is its sub-branch that learns from examples; Large language model (LLM) is a type of machine learning that specializes in language.
- Main difference: In traditional software, the human writes the rule; In machine learning, the model extracts the rule from the examples.
- AI is already in your life: spam filter, keyboard prediction, fake transaction alert, recommendation systems. What's new is that it can speak in natural language.
- Because the model works statistically, it is fluid, but it is not always accurate; this is the basis for subsequent units (hallucination, verification).
Application Task
Open an AI tool you use and explain to yourself the difference between machine learning and big language model with the "Powerful prompt" template above. Then write down 3 places in your own business where AI is already used (or could be used). For each, “What exactly is/could AI do here?” Answer the question in one sentence.
Checklist
- [ ] I can distinguish between artificial intelligence, machine learning and LLM.
- [ ] I can explain the "who writes the rules" difference between traditional software and machine learning.
- [ ] I know at least 3 examples of AI in my daily life.
- [ ] By giving context, audience, and format, I can ask a stronger question.
- [ ] I know that just because AI is smooth doesn't mean it's accurate.