AI’s Many Skills: Seeing, Hearing, Reading, Deciding

What You Will Learn

• Different types of tasks AI can do
• Examples of AI skills like seeing, listening, reading, predicting, and acting
• What AI is good at and what it is not so good at

AI as a Collection of Skills

AI is not one single magic brain.
It is more like a toolbox filled with different skills.
Each skill is trained for a specific kind of data and a specific kind of task.

Some AI skills are good with images, some with sound, some with language, and some with numbers.
Most AI systems are “narrow,” which means they are very good at one kind of thing and not very good at anything else.

AI That Sees – Computer Vision

Some AI systems are trained to work with pictures and video.
This area is called computer vision.

Computer vision can:

• Recognize faces in photos and videos
• Find objects, like cars or traffic signs, in camera images
• Help medical experts look for signs of disease in scans
• Count people or items in crowded places

These systems learn from huge collections of labeled images.
For example, many photos are marked with “cat” or “no cat,” and the AI learns patterns that match those labels.

Computer vision does not truly understand what a cat is.
It just learns that certain shapes, colors, and textures often appear together in images labeled “cat.”

AI That Listens and Speaks – Speech and Audio

AI can also work with sounds.
Speech recognition AI transforms sound into text by learning patterns between audio waves and spoken words.

This can be used to:

• Turn your voice into text on a screen
• Control devices by speaking
• Create subtitles for videos

There are also systems that go in the other direction and turn text into speech.
They can read texts out loud in a human‑like voice, sometimes even copying a specific style or accent.

Other audio AI can recognize sounds like music, clapping, or alarms, not just words.

AI That Reads and Writes – Language Models

Language models are AI systems trained on huge amounts of text.
They learn which words tend to follow other words and which sentences sound natural.

They can:

• Answer questions in text form
• Summarize long articles into shorter versions
• Translate between languages
• Generate stories, letters, or explanations

These systems do not truly understand meaning the way humans do.
They predict what text is likely to come next, based on patterns they have learned from many examples.
Because of this, they can sound confident even when they are wrong.

AI That Predicts and Recommends

Another big area of AI is prediction.
These systems look at numbers and past behavior to guess what might happen next.

Prediction AI can:

• Suggest which videos or songs you might like
• Estimate which products you might buy
• Help forecast weather and traffic patterns
• Support doctors in predicting health risks

Recommendation systems use this kind of AI to keep people engaged and interested, by choosing what to show next.

AI That Acts in the World – Robots and Agents

Some AI systems decide actions, not just labels or text.
They might control robots, cars, or characters in a game.

They can:

• Plan paths for delivery robots
• Balance and walk for robots with legs
• Control movements in factory machines
• Choose moves in complex games

These systems often learn by trying actions, getting rewards or penalties, and adjusting their behavior to get better results.

What AI Is Good At and Not Good At

AI is very good at:

• Spotting patterns in large amounts of data
• Doing the same kind of task millions of times
• Working very quickly and without getting tired

AI is not good at:

• General common sense in new and unusual situations
• Understanding feelings and context the way humans do
• Deciding what is right, fair, or kind without human guidance

Knowing these limits helps people decide where AI is safe and useful, and where human judgment is still needed.

Summary Points

• AI is a toolbox of skills, not one single magic brain
• Computer vision lets AI work with images and video
• Speech and language AI can listen, read, and generate text and speech
• Prediction AI powers recommendations and many kinds of forecasting
• Action‑focused AI can control robots, vehicles, and game agents
• AI is powerful in narrow tasks but still needs human guidance and limits

Practical Tasks

• Pick one AI skill, like image recognition or language generation, and write a few sentences describing how it works in a real app you know
• Make a simple two‑column list: “Things AI does well” and “Things humans still do better,” with at least three items in each column