メニュー

トップ

言語

初めての方へ
レッスン
講師を探す
生徒口コミ
特集
お問合せ 050-3647-0019
 
Yevi AI Demystified: How Does AI Actually Work?
IT＆プログラミング AI・ディープラーニング   流行 初心者向け 数学 理系・工学 フリートーク
 

AI Demystified: How Does AI Actually Work?

60分
5,000 ポイント
通貨別ポイント金額
Google meet icon Lattep icon
Cut through the noise — understand how artificial intelligence really works, what it's capable of today, and where the hard limits are
レッスンリクエスト
忙しい方、やり方が不明な方は 代行リクエストを依頼する

レッスンの詳細

Lesson Overview

Level: Absolute Beginner; Non-technical professional who works with AI; Technical professionals without training in ML/AI.
Duration: 60 minutes.
Format: Live 1-on-1 or small group
Goal: I will explain AI in the language that a 5th-grade pupil understands. If you're considering it as a career, it's a perfect way to see whether you want to go deeper into AI/ML, data science, or data engineering. Otherwise, it will be just very educational for you.
Teaching methods: All of these complex abstract concepts will be explained with diagrams and other images that explain the concepts in a very understandable and structured way.

Module 1 — What AI Actually Is (And Isn't) (10 min)

"AI is not a robot that thinks. It's a pattern machine that predicts."

  • The Hollywood version vs. the real version
  • A simple definition: AI is software that learns from examples instead of following fixed rules
  • The three words everyone confuses: AI, Machine Learning, Deep Learning
  • A one-slide map: how they nest inside each other
  • Real-world anchor: a spam filter is AI; so is Google Translate; so is ChatGPT — same family, very different jobs

Module 2 — How AI Actually Learns (20 min)

"You wouldn't learn to ride a bike by reading a manual. Neither does AI."

  • The core idea: training on data — show the machine millions of examples until it figures out the pattern
  • Three learning styles, in plain English:
    • Supervised learning — learning with a teacher (labeled examples)
    • Unsupervised learning — finding patterns alone (no labels)
    • Reinforcement learning — learning by trial, error, and reward
  • What a "model" is: the end result of training — a frozen set of patterns
  • Real-world anchor: teaching a child to recognize a dog vs. how image recognition AI works — same instinct, different mechanism

Module 3 — Inside a Neural Network (10 min)

"A neural network is not a brain. But it borrowed the idea from one."

  • Neurons in biology vs. nodes in AI — the analogy and where it breaks down
  • Layers explained visually: input → hidden → output
  • What "deep learning" means: just a neural network with many layers
  • How the network adjusts itself: the idea of weights and errors without the math
  • Real-world anchor: face recognition on your phone — what's actually happening when it unlocks

Module 4 — Language AI: How ChatGPT Works (10 min)

"ChatGPT doesn't know anything. It's incredibly good at sounding like it does."

  • What a Large Language Model (LLM) is in one sentence: a model trained to predict the next word, at a massive scale
  • The training process: reading a significant portion of the internet and learning patterns
  • Why it sounds so fluent: predicting the most statistically likely response
  • What a prompt is and why wording matters so much
  • The concept of hallucination — why AI confidently says wrong things
  • Real-world anchor: autocomplete on your phone keyboard, scaled up by a factor of a million

Module 5 — What AI Is Genuinely Good At (5 min)

"Give AI a pattern problem and get out of the way."

  • Where AI genuinely excels today:
    • Recognizing images, faces, speech
    • Translating languages
    • Recommending content (Netflix, Spotify, YouTube)
    • Detecting fraud and anomalies
    • Generating text, code, and images
    • Diagnosing medical images (X-rays, MRIs)
  • The common thread: all of these are pattern recognition problems
  • Real-world anchor: how Spotify knows your next favorite song before you do

Module 6 — Where AI Still Falls Short (5 min)

"AI has no common sense. And it has no idea that it doesn't."

  • The hard limits, explained plainly:
    • No true understanding — it processes patterns, not meaning
    • No common sense — it can fail on problems a five-year-old solves instantly
    • Hallucination — confident, fluent, and wrong
    • Bias — trained on human data, inherits human prejudice
    • Environment - not able to understand its environment, migrate, and function in other environments
  • What "Artificial General Intelligence (AGI)" means and why we don't have it yet
  • Real-world anchor: a famous case where an AI medical tool performed worse on darker skin tones — bias baked in by biased data

この講師の キャンセルポリシー

リクエスト確定前

  • いつでも無料キャンセル可能。

リクエスト確定後

  • レッスン開始時刻の24時間以上前→ いつでもキャンセル可能。
  • レッスン開始時刻の24時間未満→ キャンセル料がかかる場合があります。
  • レッスンに現れなかった場合→ キャンセル料がかかる場合があります。
    (キャンセル料については、講師に直接ご確認ください)

この講師のレッスン

すべてのレッスン すべて
パソコン入門
IT一般
Yevi

Yevi

出身国
居住国
1
1

レッスン可能時間帯  

スケジュールは不定期です。
リクエスト時にご確認ください。
※ Asia/Tokyo 時間で表示。
講師プロフィール
AI Demystified: How Does AI Actually Work?
60分
5,000P
Google meet icon Lattep icon
レッスンリクエスト

Yevi

出身国
 
居住国
 
1
1
講師プロフィール
お気軽にご質問ください！