1 of 13

DS161 Introduction to Data Science & Artificial Intelligence

Self-Study 1

AI Around Us: Identifying Intelligent Systems

Krishnendu Ghosh

2 of 13

Why AI Is Everywhere

From your morning alarm to bedtime streaming, AI quietly powers countless daily interactions: your music recommendations, navigation routes, spam filters, and even your phone's face unlock.

​

Stop and think: How many AI systems have you used today without even realizing it?

3 of 13

Learning Objectives

Identify AI systems in everyday environments: recognize when a product, service, or tool is powered by artificial intelligence, from recommendation engines to voice assistants and beyond.

​

Distinguish between AI and simple automation: understand the key differences between rule-based automated systems and systems that learn, adapt, or make intelligent decisions from data.

​

4 of 13

Warmup Activity

Think–Pair–Share

​

THINK: List every AI system you've used this week — phone, apps, streaming, search.

​

PAIR: Share your list with a partner. What do they have that you missed?

​

SHARE: Which examples surprised you most?

​

5 of 13

What Makes a System

"INTELLIGENT"?

Input → Perception & Representation: The system receives raw data from the environment — images, text, audio, clicks, or sensor readings. It then transforms this data into structured representations the system can work with, such as features, embeddings, or symbolic labels.

​

Reasoning & Learning: Using the structured representation, the system applies algorithms — rules, statistical models, or neural networks — to identify patterns, draw inferences, or update its internal model based on experience and feedback signals.

​

Action, Output & Feedback Loop: The system produces an output — a recommendation, decision, label, or response — and acts on it. Feedback from the outcome (user behaviour, error signals, rewards) flows back into the system, enabling continuous improvement over time.

6 of 13

AI or just Automation?

​

🌡 Thermostat (Automation)

Follows fixed rules: "If temp < 68°F, turn on heat." No learning, no adaptation. The same rule runs every time — regardless of your habits or preferences. Rule-based. Predictable. Not AI.

​

​

🧠 Smart Thermostat (AI)

Learns your daily schedule, adapts to your preferences, and predicts when you'll be home. Improves over time with data. Learning-based. Adaptive. This is AI.

​

7 of 13

AI Around Us

​

Recommendations & Navigation: Streaming platforms like Netflix suggest content based on your watch history. Google Maps predicts traffic and reroutes you in real time using live data from millions of users.

​

​

Voice Assistants, Spam Filters & Vision: Siri and Alexa understand natural language to respond and act. Email spam filters learn to block unwanted messages. Face recognition unlocks phones and tags photos; smart cameras detect objects and motion automatically.

​

8 of 13

AI Capability Map

​

​

Perception & Language Understanding: AI systems can interpret the world through inputs like images, audio, and text. Examples include computer vision (detecting objects in photos), speech recognition (converting spoken words to text), and natural language processing (understanding and generating human language in chatbots or translators).

​

Prediction, Reasoning & Generation: AI systems learn from data to forecast outcomes, make decisions, and create new content. Examples include recommendation engines (predicting what you will watch next), decision support systems (diagnosing disease from symptoms), and generative models (producing realistic images, code, or written text from a prompt).

9 of 13

Activity: Identify AI Capability

Match each system to its primary AI capability:

  • Autocorrect → ?
  • Fraud Detection → ?
  • Chatbots → ?

Choices: Natural Language Processing · Anomaly Detection · Sequence Prediction

​

​

Match each system to its primary AI capability:

  • Self-Driving Cars → ?
  • Recommendation Engines → ?
  • Image Search → ?

Choices: Computer Vision · Collaborative Filtering · Perception & Decision-Making

10 of 13

Mini Case Study

​

YouTube & Netflix Recommendation Systems

​

What data? Viewing history, ratings, watch time, search queries, and user demographics are collected continuously.

What prediction? The system predicts which content you are most likely to watch next and enjoy based on patterns from millions of users.

What action? A personalized feed is generated — thumbnails, autoplay suggestions, and homepages are all uniquely tailored to each user in real time.

11 of 13

Critical Thinking

Is every system using an algorithm actually AI?

​

Ask yourself: Does it learn from data? Does it adapt to new inputs? Does it handle uncertainty? A spam filter with fixed rules differs from one that updates itself. What evidence would convince you a system is truly intelligent?

12 of 13

Reflection + Take-Home Task

Think: Where did you encounter AI today — before this class even started?

​

Take-Home Task: Identify 3 systems you interact with regularly. For each system, describe the following:

  • Input Data — What information does the system collect or receive? (e.g., your clicks, location, voice, history)
  • AI Capability Used — What type of AI is at work? (e.g., recommendation, perception, prediction, natural language understanding)
  • Output / Action — What does the system do or decide as a result?
  • Is AI Actually Necessary? — Could simpler rule-based automation achieve the same result? What evidence supports your view?

13 of 13