After this chapter you will be able to
- explain cloud computing, its service and deployment models
- distinguish AI, machine learning, deep learning and generative AI
- describe IoT, blockchain, AR/VR/MR, quantum computing and robotics
- explain big data, edge computing and other new terms
- recall India’s missions and initiatives in these areas
34.1 Cloud Computing
Cloud computing is the delivery of computing services — servers, storage, databases, software, AI — over the internet, on demand, pay-as-you-go.
Key features (NIST): on-demand self-service · broad network access · resource pooling · rapid elasticity (scale up/down) · measured service (pay for use).
Service Models
| Layer | On-premises | IaaS | PaaS | SaaS |
|---|---|---|---|---|
| Applications | You | You | You | Provider |
| Data | You | You | You | Provider |
| Runtime | You | You | Provider | Provider |
| Middleware | You | You | Provider | Provider |
| OS | You | You | Provider | Provider |
| Virtualisation | You | Provider | Provider | Provider |
| Servers | You | Provider | Provider | Provider |
| Storage | You | Provider | Provider | Provider |
| Networking | You | Provider | Provider | Provider |
Examples: own data centre · AWS EC2 (IaaS) · Google App Engine (PaaS) · Gmail / Microsoft 365 (SaaS).
| Model | Full form | You get | Examples |
|---|---|---|---|
| IaaS | Infrastructure as a Service | Virtual machines, storage, networks — you manage OS and apps | Amazon EC2 (AWS), Microsoft Azure VMs, Google Compute Engine, DigitalOcean, Linode, Rackspace |
| PaaS | Platform as a Service | A ready platform to build and run apps | Google App Engine, Heroku, Azure App Service, AWS Elastic Beanstalk, Force.com |
| SaaS | Software as a Service | Ready software used through a browser | Gmail, Google Docs, Microsoft 365, Zoom, Salesforce, Dropbox |
Other “as-a-Service” terms: FaaS (serverless functions — AWS Lambda), DaaS (Desktop), STaaS (Storage), AIaaS.
Deployment Models
| Model | Meaning |
|---|---|
| Public cloud | Shared infrastructure owned by a provider, open to everyone (AWS, Azure, Google Cloud, IBM Cloud, Oracle Cloud) |
| Private cloud | Used by one organisation only |
| Hybrid cloud | Mix of public and private |
| Community cloud | Shared by organisations with common needs (e.g., government departments) |
| Multi-cloud | Using services from several providers |
- Largest cloud providers: Amazon Web Services (AWS), Microsoft Azure, Google Cloud.
- Hybrid cloud examples: AWS Outposts, Azure Stack, Azure Arc and Google Anthos.
- India: MeghRaj — the Government of India’s cloud initiative (GI Cloud); government-empanelled Indian data centres.
- Virtualisation (running many virtual machines on one physical server) is the technology behind the cloud.
- Everyday cloud use: Google Drive or iCloud backup, editing in Google Docs, and streaming on Netflix.
- Popular cloud storage services: Google Drive, Microsoft OneDrive, Dropbox, iCloud and Box. McAfee is an antivirus and security company, not a cloud storage service.
★ EXAM POINT — SaaS = software over the internet (Gmail, Microsoft 365); PaaS = platform for developers; IaaS = raw infrastructure (virtual servers). The term “cloud computing” was popularised in 2006 (Amazon EC2 launch; Google’s Eric Schmidt).
34.2 Artificial Intelligence (AI)
Artificial Intelligence is the ability of computers to perform tasks that normally need human intelligence — learning, reasoning, understanding language, recognising images, making decisions.
- The term was coined by John McCarthy at the Dartmouth Conference (1956) — “father of AI”.
- John McCarthy (1927–2011) also developed the LISP language and won the Turing Award (1971).
- AI is also called Machine Intelligence.
- Alan Turing proposed the Turing Test (1950) to judge machine intelligence. His 1950 paper was titled Computing Machinery and Intelligence.
- Alan Turing is widely called the father of computer science. He defined the Turing machine (1936), a model of general computation.
| Term | Meaning | Example |
|---|---|---|
| Machine Learning (ML) | Systems learn patterns from data instead of being explicitly programmed (term by Arthur Samuel, 1959) | Spam filters, recommendations |
| Supervised learning | Learns from labelled examples | Predicting exam scores |
| Unsupervised learning | Finds patterns in unlabelled data (clustering) | Customer segments |
| Reinforcement learning | Learns by reward and penalty | Game-playing AI, robots |
| Deep Learning | ML using artificial neural networks with many layers | Face recognition, speech |
| Neural network | Layers of connected nodes inspired by the brain | — |
| NLP | Natural Language Processing — understanding human language | Translation, chatbots, voice assistants |
| Computer vision | Understanding images and video | Number-plate reading, medical scans |
| Generative AI (GenAI) | Creates new content — text, images, audio, video, code | ChatGPT, Gemini, Copilot, Claude, image generators |
| LLM | Large Language Model — AI trained on huge text data to understand and generate language | GPT, Gemini, Llama, Claude |
| Prompt | The instruction given to a generative AI | — |
| Chatbot | Program that chats with users | Bank help-bots, ChatGPT (OpenAI, Nov 2022) |
| Expert system | AI that gives advice in one field using rules | Medical diagnosis systems (MYCIN) |
| Hallucination | AI confidently giving wrong or made-up information | — |
| Deepfake | AI-generated fake video/audio of real people | — |
| Type of AI | Meaning |
|---|---|
| Narrow (weak) AI | Good at one task — all AI today |
| General AI (AGI) | Human-level intelligence across tasks — not yet achieved |
| Super AI | Beyond human intelligence — hypothetical |
Famous AI milestones: IBM Deep Blue beat chess champion Garry Kasparov (1997) · IBM Watson won Jeopardy! (2011) · Google DeepMind’s AlphaGo beat Lee Sedol at Go (2016) · ChatGPT launched (2022) · Sophia robot given Saudi citizenship (2017).
Voice assistants: Siri (Apple), Alexa (Amazon), Google Assistant/Gemini, Bixby (Samsung), Cortana (Microsoft, retired).
Launch years: Siri (2011), Google Now (2012, later became Google Assistant in 2016), Alexa and Cortana (2014), Bixby (2017).
India and AI: IndiaAI Mission (approved March 2024, outlay about ₹10,372 crore) — compute capacity, datasets, Indian foundation models; AIRAWAT AI supercomputer; Bhashini for Indian-language AI; BharatGen (government-supported Indian LLM initiative); India hosts the AI Impact Summit (New Delhi, February 2026).
AI Programming Languages
| Language | Note |
|---|---|
| Python | Most widely used language for AI and machine learning |
| LISP (List Processing) | Created by John McCarthy; used in early AI research |
| Prolog (Programming in Logic) | Logic-based language; created by Alain Colmerauer |
Where AI Is Used
| Field | Example use |
|---|---|
| Business | Automating repetitive office tasks (RPA) |
| Banking | Customer-support chatbots; detecting anomalies and credit card fraud |
| Healthcare | Faster diagnosis; IBM Watson understands natural language |
| E-commerce | Product recommendations, dynamic pricing |
| Gaming | Intelligent opponents; one of AI’s biggest success areas |
| Transport | Self-driving cars and route optimisation |
| Agriculture | Crop disease detection, precision farming |
| Education | Personalised learning, automated grading |
How a Machine Learning Model Is Built
Typical steps: data collection → cleaning → splitting → training → evaluation → deployment → monitoring.
| Data set | Share (approx.) | Purpose |
|---|---|---|
| Training set | 70–80% | Teaches the model |
| Validation set | 10–15% | Tunes the model |
| Test set | 10–15% | Checks performance on unseen data |