Learning Objectives
- Define artificial intelligence in IT 231 terms (not a research paper).
- Identify the main types and applications of AI.
- Describe impacts on business (productivity, new products, risk) and society (jobs, bias, access).
- Explain the principles of responsible AI use in organizations.
- Place AI next to other Unit 10 topics (ML, cloud, IoT) without replacing them.
This chapter opens Unit 10. Later notes cover data science, ML, and neural networks, cloud, green, and virtual computing, big data and blockchain, social media and digital marketing, and IoT.
What AI is here
Artificial intelligence is the field of making computers perform tasks that usually require human intelligence: recognizing speech and images, recommending products, translating language, detecting fraud, or drafting text. Narrow AI (today’s systems) is good at specific tasks. General AI that matches humans across all tasks is not what current business tools are.
Machine learning (next chapter) is a common way to build AI: the system improves from data rather than only from hand-written rules. Neural networks are one ML technique.
Main branches and applications
| Branch | What it does | Business example |
|---|---|---|
| Natural language processing (NLP) | Understands and generates human language | Chatbots, translation, summarizing documents |
| Computer vision | Interprets images and video | Face login, quality inspection, number-plate reading |
| Speech recognition | Converts speech to text and back | Voice assistants, call-center transcription |
| Recommendation systems | Suggests items based on behaviour | “Customers also bought”, streaming suggestions |
| Predictive analytics | Forecasts outcomes from past data | Sales forecasting, credit scoring, churn prediction |
| Generative AI | Creates new text, images, audio, or code | Drafting marketing copy, product images, code help |
| Robotics and automation | Physical or software actions with AI guidance | Warehouse robots, automated invoice processing |
Rule-based software vs AI
A traditional program follows rules written by a programmer: “if the transaction is over Rs 5 lakh, flag it.” An AI fraud model learns patterns from thousands of past fraud cases, and can flag an unusual Rs 20,000 transaction at 3 a.m. from a new device that a fixed rule would miss. The trade-off is that its reasoning is harder to explain.
Impact on business
- Productivity: automation of routine classification, customer chat, document drafting, and forecasting.
- New products: recommendation engines, credit scoring, route planning, personalized marketing.
- Cost and skill mix: fewer hours on repetitive work; more need for people who can check outputs and manage data.
- Risks: wrong decisions at scale, privacy, dependency on vendors, and “hallucinated” facts in generative tools.
Managers should treat AI as a tool inside an information system: data quality, process design, and accountability still matter.
| Business function | AI application |
|---|---|
| Marketing | Customer segmentation, ad targeting, content drafting |
| Sales and service | Chatbots for FAQs, lead scoring |
| Finance | Fraud detection, credit scoring, invoice processing |
| HR | Screening CVs (with bias checks), answering staff queries |
| Operations | Demand forecasting, route optimization, predictive maintenance |
Impact on society
- Jobs: some roles shrink (data entry, basic translation); others grow (AI-supported analysis, oversight, prompt and process design). Displacement is uneven by sector and education.
- Access: language tools can help Nepali users, but models trained mainly on English data can perform worse on local languages and contexts.
- Fairness: biased training data can reproduce discrimination in hiring or credit.
- Public life: deepfakes, misinformation, and surveillance concerns require digital literacy (Unit 9) as well as technology.
- Healthcare and education: AI can help diagnose diseases from scans and personalize learning, extending scarce expertise to remote areas.
flowchart TB
AI[Artificial Intelligence]
AI --> B[Business Impact]
AI --> S[Society Impact]
B --> P[Productivity and automation]
B --> N[New products and services]
B --> R[Risks: bias privacy vendor lock-in]
S --> J[Jobs: displacement and new roles]
S --> A[Access and language fairness]
S --> F[Fairness and public trust]
Figure 1: AI Impact on Business and Society
Benefits and concerns
| Benefits | Concerns |
|---|---|
| Faster, cheaper services available 24/7 | Job displacement in routine roles |
| Better decisions from data | Bias and unfair outcomes |
| New products and industries | Privacy and surveillance |
| Help with language, disability, and access | Misinformation and deepfakes |
| Support for doctors, teachers, and farmers | Over-reliance on systems that can be confidently wrong |
Responsible AI in an organization
- Human oversight — people review important decisions (loans, hiring, medical).
- Transparency — tell customers when they are dealing with AI, such as a chatbot.
- Fairness checks — test whether outcomes differ unfairly by gender, caste, region, or age.
- Privacy — use only the data needed, with consent and security.
- Accuracy checks — verify generative AI output before publishing or sending.
- Accountability — a named person or team is responsible for each AI system.
Nepal example
Banks, wallets, and telecoms already use scoring and fraud rules that are AI-adjacent. A shop using a chatbot for FAQs still needs a human path for complaints. Policy and Electronic Transactions Act issues (Unit 9) apply to automated decisions that affect customers.
Other examples include AI-assisted crop disease identification from phone photos for farmers, Nepali-language speech and translation tools, and students and freelancers using generative AI for drafting and coding. Nepal’s government has also begun work on a national AI policy framework, reflecting growing attention to how AI should be governed.
Key Takeaways
- Course-level AI is narrow, data-driven automation and assistance—not science fiction.
- AI branches include NLP, computer vision, recommendation, prediction, and generative AI.
- Business gains come with operational and ethical risk.
- Society feels AI through jobs, language access, bias, and information quality.
- Responsible AI needs human oversight, transparency, fairness, privacy, and accountability.
Discussion Questions
- Name one task in a Nepali SME that is a good AI candidate and one that is not.
- Why is “the model said so” a weak management control?
- How does AI depend on cloud computing and big data without being the same thing?
Exam-Style Questions
Short answer
- Define artificial intelligence. Differentiate between narrow AI and general AI.
- What is generative AI? Give two business uses.
- List four principles of responsible AI.
Long answer
- Discuss the impact of AI on business and society, with examples.
- A Nepali bank wants to use AI for loan approval. Explain the benefits, risks, and safeguards it should consider.
