Artificial intelligence
| English | Chinese | Pinyin |
|---|---|---|
| artificial intelligence | 人工智能 | rén gōng zhì néng |
| machine learning | 机器学习 | jī qì xué xí |
| deep learning | 深度学习 | shēn dù xué xí |
| neural networks | 神经网络 | shén jīng wǎng luò |
| training data | 训练数据 | xùn liàn shù jù |
| bias | 偏见 | piān jiàn |
| accountability | 问责 | wèn zé |
Machines that learn
- Artificial intelligence 人工智能 (AI) builds systems that do tasks once thought to need human intelligence — recognising speech and images, translating, driving.
- Modern AI is powerful and everywhere, but it raises real concerns.
- Let's see how it works, what it does, and where the risks lie.
Computing concept lab
Classify concrete examples by the computing idea they demonstrate.
How modern AI works
- Most AI uses machine learning 机器学习 — algorithms that improve at a task by learning patterns from large amounts of data, instead of being programmed step by step.
- Deep learning 深度学习, using neural networks 神经网络 with many layers, is the leading approach today.
- The more (and better) the training data 训练数据, the better the model — but bad data leads to bad results.
A neural network learns patterns from data through layers of connected nodes
Machine learning differs from traditional programming because it:
ML algorithms improve at a task by learning from large amounts of data, rather than following hand-written rules.
Match each AI term to what it means.
ML learns from data; deep learning uses many layers; bias and opacity (black box) are key ethical concerns.
Everyday examples
- Speech recognition (voice assistants) and image recognition (faces, objects, text).
- Machine translation between languages; recommendation systems (videos, products).
- Autonomous vehicles and robots.
- A classic combo: a camera reads a label with OCR, machine translation converts the text, and text-to-speech reads it aloud.
Which is an everyday application of AI?
Recommendation systems (and speech/image recognition, translation, self-driving) are common AI applications.
Benefits
- Accessibility — speech/image AI helps users with impairments; translation helps non-native speakers.
- Productivity — automating repetitive tasks frees people for creative work.
- Decision support — AI spots patterns in huge datasets (medical diagnosis, fraud detection).
- It is always available and personalised to each user.
A benefit of AI is:
AI excels at finding patterns in large data (decision support). Fairness and explainability are concerns, not guarantees.
Concerns
- Bias 偏见 — unfair patterns in the training data become unfair AI decisions (hiring, lending).
- Job displacement, privacy (training on personal data), and transparency (big models are "black boxes").
- Accountability 问责 — when AI is wrong, who is responsible: developer, user, or operator?
- Misuse — deepfakes, misinformation, surveillance. Professionals must know the limits, inform users, and reduce harm.
AI "bias" usually arises because:
If the training data reflects unfair patterns, the model learns and repeats them — e.g. in hiring or lending.
Select all genuine concerns about AI.
Bias, opacity and misuse are real concerns. AI certainly can be wrong — accountability for that is itself a concern.
You've got it
- machine learning improves from data, not step-by-step programming; deep learning uses neural networks
- examples: speech/image recognition, translation, recommendations, self-driving
- benefits: accessibility, productivity, decision support
- concerns: bias, job loss, privacy, transparency, accountability, misuse