🧠 1. What is Artificial Intelligence?
Definition:
👉 Artificial Intelligence (AI) means machines (computers) performing tasks that normally require human intelligence, such as learning, reasoning, problem-solving, understanding language, and making decisions.
In simple words:
AI enables computers to think, learn, and act like humans.
🔍 Examples of AI in daily life:
| Area | Example |
|---|---|
| Smartphones | Voice assistants like Siri, Google Assistant |
| Banking | Fraud detection, credit scoring, chatbots |
| E-commerce | Product recommendations (Amazon, Flipkart) |
| Transport | Google Maps traffic prediction |
| Finance | Algorithmic trading, robo-advisors |
| Security | Face recognition, biometric verification |
⚙️ 2. Key Features of AI
- Learning: Systems learn from data (experience).
- Reasoning: Make decisions using logic and rules.
- Problem-Solving: Find best solutions automatically.
- Perception: Understand inputs (images, voice, text).
- Adaptability: Improve with experience (self-learning).
🧩 3. Types of Artificial Intelligence (Based on Capability)
| Type | Description | Example |
|---|---|---|
| Narrow AI (Weak AI) | AI designed for one specific task only. | Chatbots, recommendation systems |
| General AI (Strong AI) | Can perform any intellectual task like a human (still theoretical). | Not yet achieved |
| Super AI | AI that surpasses human intelligence (future concept). | Imaginary, in research stage |
👉 Today’s AI is Narrow AI — built for specific applications only.
💡 4. Types of AI (Based on Functionality)
| Type | Description | Example |
|---|---|---|
| Reactive Machines | Only react to situations; no memory. | IBM Deep Blue (chess computer) |
| Limited Memory | Learn from past data to improve decisions. | Self-driving cars, chatbots |
| Theory of Mind | Can understand human emotions & beliefs. | Not yet fully developed |
| Self-aware AI | Has consciousness and emotions. | Theoretical / future concept |
🧮 5. Subfields / Branches of AI
| Field | Meaning | Example |
|---|---|---|
| Machine Learning (ML) | Computers learn automatically from data without explicit programming. | Spam detection, credit risk models |
| Deep Learning | Advanced ML using neural networks for large data & complex tasks. | Voice & image recognition |
| Natural Language Processing (NLP) | Helps computers understand and respond to human language. | Chatbots, Google Translate |
| Computer Vision | Enables computers to “see” and interpret images/videos. | Face unlock, CCTV analytics |
| Robotics | AI + mechanical systems to perform tasks. | Industrial robots, drones |
| Expert Systems | Mimic decisions of human experts using rule-based knowledge. | Medical diagnosis system |
| Speech Recognition | Converts spoken words to text. | Alexa, Google Assistant |
🏦 6. Applications of AI in Banking & Financial Sector
| Area | Use Case | Example |
|---|---|---|
| Customer Service | 24×7 AI chatbots handle queries. | HDFC EVA chatbot |
| Fraud Detection | Detects unusual transaction patterns. | SBI Card fraud alerts |
| Credit Scoring | AI models evaluate loan eligibility. | Automated risk scoring systems |
| Robo-Advisors | Provide personalized investment advice. | Zerodha Varsity, Upstox AI |
| Document Processing | AI scans & extracts info from forms. | KYC, loan applications |
| Cybersecurity | AI monitors suspicious activities. | Fraud prevention tools |
| Process Automation | Repetitive tasks automated via RPA (Robotic Process Automation). | Account reconciliation, report generation |
| Predictive Analytics | Predicts loan default or churn. | Risk modeling systems |
🧠 7. How AI Works (Basic Idea)
AI systems follow 3 main steps:
- Input → Data collected (images, text, numbers, etc.)
- Processing → Algorithms analyze patterns & learn.
- Output → AI makes prediction or decision.
Example:
- Bank feeds transaction data → AI detects fraud pattern → flags suspicious activity.
🔢 8. Difference between AI, ML, and Deep Learning
| Concept | What It Does | Example |
|---|---|---|
| AI | Simulates human intelligence | Chatbot, self-driving car |
| Machine Learning (ML) | System learns from data | Credit scoring, spam filtering |
| Deep Learning (DL) | ML using neural networks (large data) | Face recognition, voice assistants |
Remember:
AI → ML → DL
(Deep Learning is a subset of Machine Learning, which is a subset of AI)
🔐 9. Advantages of AI
- Efficiency: Performs repetitive tasks quickly.
- Accuracy: Fewer human errors.
- Availability: Works 24×7.
- Cost-saving: Reduces manpower for routine jobs.
- Data-driven: Better decisions using analytics.
- Personalization: Tailored customer experience.
⚠️ 10. Limitations / Challenges of AI
- High cost of setup and maintenance.
- Data privacy & security concerns.
- Job displacement — automation replaces some human roles.
- Bias in AI models (if training data is biased).
- Lack of transparency (black-box algorithms).
- Regulatory & ethical concerns (RBI, SEBI oversight).
🧩 11. AI vs Human Intelligence
| Aspect | Artificial Intelligence | Human Intelligence |
|---|---|---|
| Speed | Very fast in processing data | Slower |
| Creativity | Limited | High creativity |
| Learning | Needs data to learn | Learns from experience & emotions |
| Adaptability | Depends on programming | Flexible & self-adaptive |
| Emotion | No emotions | Emotional & social understanding |
🏛️ 12. AI in Government & Regulation
- NITI Aayog launched National Strategy for Artificial Intelligence (AI for All).
- Focus areas: Healthcare, Agriculture, Education, Smart Cities, and FinTech.
- RBI & SEBI encourage use of AI/ML for financial inclusion, fraud detection, and market surveillance.
- NABARD explores AI for agriculture risk assessment.
⚙️ 13. Real-life Examples (Banking & Beyond)
- SBI: YONO app uses AI for product recommendations.
- HDFC Bank: EVA chatbot answers customer queries.
- ICICI Bank: Uses RPA for 200+ processes.
- Axis Bank: “Aha!” AI chatbot for banking support.
- Google Pay / PhonePe: AI detects fraudulent payments.
- Insurance sector: AI for claim automation and fraud checks.
🔍 14. Future of AI in Banking
- Smarter robo-advisors for personal finance.
- AI-based credit assessment for MSMEs.
- Voice-enabled banking (speech-to-text transactions).
- Predictive systems for loan defaults and market trends.
- Explainable AI (XAI) for transparency in decision-making.
🧾 15. Common Terms (for quick revision)
| Term | Meaning |
|---|---|
| Algorithm | Step-by-step rules for solving a problem. |
| Big Data | Huge data sets used for AI analysis. |
| Neural Network | AI system modeled after the human brain. |
| Chatbot | AI program that simulates conversation. |
| Automation | Performing tasks without human intervention. |
| RPA (Robotic Process Automation) | Automating repetitive office tasks using bots. |
| Predictive Analytics | Using data to forecast future trends. |
| Cognitive Computing | AI that mimics human thought processes. |
🧭 16. Quick Revision Summary Table
| Concept | Key Point |
|---|---|
| Definition | Machines that think and act like humans |
| Current Type | Narrow AI (task-specific) |
| Branches | ML, DL, NLP, Vision, Robotics, Expert Systems |
| Banking Uses | Fraud detection, Chatbots, Credit scoring, RPA |
| Govt Initiative | NITI Aayog – AI for All |
| Benefits | Speed, Accuracy, 24×7, Cost-saving |
| Challenges | Cost, Bias, Job loss, Privacy, Ethics |
🧠 17. One-Liner Exam Triggers
- AI = Intelligence by machines.
- ML = AI that learns from data.
- NITI Aayog’s theme: “AI for All”.
- India’s AI focus areas: Health, Agri, Edu, Smart Cities, FinTech.
- In banks: AI used for fraud detection, RPA, chatbots, customer analytics.
- RBI: Promotes ethical & transparent AI in finance.
- Example: HDFC EVA chatbot → Narrow AI.
- Limitation: Data bias and high cost.
Exam Specific Questions
2019 IBPS IT Officer Exam:
- What is the primary goal of Artificial Intelligence?
- Answer: To create intelligent machines.
- What is the difference between Machine Learning and Deep Learning?
- Answer: Machine Learning is a broader concept that includes various algorithms that allow computers to learn from data. Deep Learning is a subset of Machine Learning that uses neural networks with many layers (deep networks) to analyze various factors of data.
- What is the term for a machine’s ability to learn from data without being explicitly programmed?
- Answer: Machine Learning.
2018 IBPS IT Officer Exam:
- What is the primary application of Natural Language Processing (NLP)?
- Answer: To enable computers to understand, interpret, and respond to human language in a valuable way.
- What is the difference between Supervised and Unsupervised Learning?
- Answer: Supervised Learning uses labeled data to train models, while Unsupervised Learning uses unlabeled data to find patterns or groupings.
- What is the term for a type of neural network that is used for image recognition?
- Answer: Convolutional Neural Network (CNN).
2017 IBPS IT Officer Exam:
- What is the primary goal of Robotics?
- Answer: To enable computers and machines to interact with the physical world and perform tasks autonomously.
- What is the term for a type of machine learning algorithm that is used for classification?
- Answer: Classification algorithms (e.g., Decision Trees, Support Vector Machines, etc.).
- What is the difference between a Neural Network and a Deep Neural Network?
- Answer: A Neural Network typically has one or two layers, while a Deep Neural Network has multiple layers (deep architecture) that allow it to learn complex patterns.
2016 IBPS IT Officer Exam:
Answer: A Decision Tree is a single tree structure used for making decisions based on feature values, while a Random Forest is an ensemble of multiple decision trees that improves accuracy and reduces overfitting.
What is the primary application of Computer Vision?
Answer: To enable machines to interpret and make decisions based on visual data from the world, such as images and videos.
What is the term for a type of machine learning algorithm that is used for regression?
Answer: Regression algorithms (e.g., Linear Regression, Polynomial Regression, etc.).
What is the difference between a Decision Tree and a Random Forest?
Answer: A Decision Tree is a single tree structure used for making decisions based on feature values, while a Random Forest is an ensemble of multiple decision trees that improves accuracy and reduces overfitting.
CHAPTER 1: BASICS OF ARTIFICIAL INTELLIGENCE (10 MCQs)
Q1. Artificial Intelligence (AI) mainly refers to:
a) Machines doing only calculations
b) Machines that can think and act like humans to some extent
c) Only robots moving physically
d) Only storing large data
Answer: b) Machines that can think and act like humans to some extent
Explanation: AI enables machines to mimic human-like intelligence such as learning, reasoning and decision making. 👉 (HIGHLY IMPORTANT)
Q2. Which of the following is the BEST example of AI?
a) Manual ledger posting
b) Calculator performing addition
c) Chatbot answering customer queries
d) Photocopy machine
Answer: c) Chatbot answering customer queries
Explanation: Chatbots use AI (NLP + ML) to understand and respond to customers.
Q3. AI that is designed for a specific task only is called:
a) General AI
b) Strong AI
c) Narrow AI
d) Super AI
Answer: c) Narrow AI
Explanation: Narrow AI is trained for one specific task like face recognition or language translation.
Q4. Which of the following is NOT a goal of AI?
a) Learning from data
b) Decision making
c) Pattern recognition
d) Increasing manual paperwork
Answer: d) Increasing manual paperwork
Explanation: AI aims to automate and reduce manual work, not increase it.
Q5. Machine Learning (ML) is a branch of AI that:
a) Hard-codes all rules manually
b) Allows systems to learn from data and improve performance
c) Works only on mechanical devices
d) Can work without any data
Answer: b) Allows systems to learn from data and improve performance
Explanation: ML models learn from past data patterns. 👉 (HIGHLY IMPORTANT)
Q6. Which of the following is MOST related to AI?
a) Internet bandwidth
b) Natural Language Processing (NLP)
c) Power backup
d) Office furniture
Answer: b) Natural Language Processing (NLP)
Explanation: NLP is an AI technique to understand and process human language.
Q7. AI system that can perform any intellectual task that a human can do is called:
a) Narrow AI
b) General AI
c) Reactive AI
d) Static AI
Answer: b) General AI
Explanation: General AI refers to human-level intelligence across many tasks (still theoretical).
Q8. Training an AI model means:
a) Paying salary to the model
b) Feeding data and adjusting model parameters
c) Changing hardware only
d) Printing reports
Answer: b) Feeding data and adjusting model parameters
Explanation: Models learn patterns during training using data.
Q9. Which term is closely associated with AI-based decision making?
a) Heuristic and learning
b) Only manual sampling
c) Punch cards
d) Typewriting
Answer: a) Heuristic and learning
Explanation: AI uses heuristics and learning algorithms to make decisions.
Q10. The main requirement for AI/ML systems is:
a) Large amount of quality data
b) Manual registers
c) Fixed interest rates
d) Extra office space
Answer: a) Large amount of quality data
Explanation: Data is the key fuel for AI systems. 👉 (HIGHLY IMPORTANT)
CHAPTER 2: ADVANCED CONCEPTS & TERMINOLOGY (15 MCQs)
Q11. Supervised learning in AI requires:
a) No data labels
b) Only numeric data
c) Input data with correct output labels
d) Only images
Answer: c) Input data with correct output labels
Explanation: In supervised learning, models learn from labelled examples.
Q12. Unsupervised learning is mainly used for:
a) Classification
b) Regression
c) Clustering and pattern discovery
d) Payroll processing
Answer: c) Clustering and pattern discovery
Explanation: It groups similar data without labels.
Q13. Which of the following is an example of supervised learning in banking?
a) Grouping customers with similar spending patterns without labels
b) Predicting loan default based on past labelled data
c) Random filing of customer forms
d) Counting cash manually
Answer: b) Predicting loan default based on past labelled data
Explanation: Here, past default/non-default labels are used to train the model.
Q14. A neural network is inspired by:
a) Human nervous system and brain
b) Car engine
c) Hydraulic pump
d) Filing cabinet
Answer: a) Human nervous system and brain
Explanation: Neural networks are modeled loosely on neurons and their connections.
Q15. Overfitting in AI models means:
a) Model performs well on training data but poorly on new data
b) Model is too simple
c) Model using less data
d) Model not trained at all
Answer: a) Model performs well on training data but poorly on new data
Explanation: Overfitting is a key risk in ML models. 👉 (HIGHLY IMPORTANT)
Q16. Which AI term is related to giving machines the ability to “see” and interpret images?
a) NLP
b) Computer Vision
c) Robotics
d) Cryptography
Answer: b) Computer Vision
Explanation: It deals with processing and understanding visual information.
Q17. Chatbots mainly use which branch of AI?
a) Robotics
b) NLP and ML
c) Cyber security
d) Image compression
Answer: b) NLP and ML
Explanation: Chatbots understand language and learn from interactions.
Q18. A model used to predict whether a transaction is fraud or not is an example of:
a) Clustering
b) Classification
c) Sorting
d) Sampling
Answer: b) Classification
Explanation: Classification assigns items to categories like fraud/non-fraud.
Q19. In AI, the term “algorithm” means:
a) Software license
b) Step-by-step method or rule for solving a problem
c) Hardware device
d) Cloud provider
Answer: b) Step-by-step method or rule for solving a problem
Explanation: Algorithms define how data is processed.
Q20. Bias in AI models arises mainly due to:
a) Only high hardware speed
b) Unrepresentative or unfair training data
c) Low internet speed
d) Manual authorization
Answer: b) Unrepresentative or unfair training data
Explanation: Biased data leads to biased outcomes. 👉 (HIGHLY IMPORTANT)
Q21. Deep Learning is a subset of:
a) Cloud computing
b) Machine Learning
c) Manual testing
d) Cryptography
Answer: b) Machine Learning
Explanation: Deep learning uses multi-layer neural networks.
Q22. Which AI technique is MOST suitable for sentiment analysis of customer feedback?
a) Robotics
b) NLP with ML
c) Clustering hardware
d) Optical storage
Answer: b) NLP with ML
Explanation: NLP processes text to find sentiment.
Q23. “Black box” in AI refers to:
a) A locked room in bank
b) A model whose internal logic is not easily interpretable
c) ATM safe
d) CCTV recording device
Answer: b) A model whose internal logic is not easily interpretable
Explanation: Many complex models lack explainability.
Q24. Model “drift” in AI means:
a) Staff transfer
b) Gradual decline in model performance over time
c) Power fluctuation
d) Network shifting
Answer: b) Gradual decline in model performance over time
Explanation: Due to change in data patterns over time.
Q25. In banking AI systems, “explainability” is important because:
a) It increases power consumption
b) Regulators and customers need to understand key reasons for decisions
c) It hides decisions
d) It removes transparency
Answer: b) Regulators and customers need to understand key reasons for decisions
Explanation: Especially important in credit and risk decisions.
CHAPTER 3: APPLICATIONS OF AI IN BANKING & FINANCE (15 MCQs)
Q26. AI-based credit scoring models help banks to:
a) Decide branch location only
b) Assess borrower risk using multiple variables
c) Only print passbooks
d) Manage lockers
Answer: b) Assess borrower risk using multiple variables
Explanation: AI improves quality of credit decisions. 👉 (HIGHLY IMPORTANT)
Q27. In banking, AI chatbots are mainly used for:
a) Physical security
b) Handling customer queries 24×7
c) Filling loan forms manually
d) Counting cash in vaults
Answer: b) Handling customer queries 24×7
Explanation: Chatbots reduce workload on call centres.
Q28. AI-based fraud detection systems primarily look for:
a) Random transactions
b) Patterns and anomalies in transaction data
c) Cheque size only
d) Branch code only
Answer: b) Patterns and anomalies in transaction data
Explanation: Unusual behaviour is flagged as potential fraud. 👉 (HIGHLY IMPORTANT)
Q29. AI in AML (Anti-Money Laundering) helps to:
a) Increase manual file checks
b) Monitor large volumes of transactions and detect suspicious activity
c) Print KYC forms
d) Approve all transactions blindly
Answer: b) Monitor large volumes of transactions and detect suspicious activity
Explanation: AI supports automated AML monitoring.
Q30. Robo-advisors in wealth management use AI to:
a) Allocate funds to random assets
b) Suggest personalized investment portfolios based on customer profile
c) Store gold in lockers
d) Decide office timings
Answer: b) Suggest personalized investment portfolios based on customer profile
Explanation: AI-based advisors give automated advice.
Q31. Which is the BEST example of AI in customer service?
a) Static FAQ PDF
b) IVR menu without understanding speech
c) Intelligent virtual assistant understanding natural language
d) Printed brochure
Answer: c) Intelligent virtual assistant understanding natural language
Explanation: It uses NLP and ML to assist users.
Q32. AI can help in loan collection by:
a) Ignoring overdue accounts
b) Predicting default risk and suggesting recovery strategy
c) Removing all contact data
d) Closing branches
Answer: b) Predicting default risk and suggesting recovery strategy
Explanation: Predictive analytics supports collection management.
Q33. In risk management, AI helps by:
a) Hiding risk
b) Providing early warning signals based on patterns
c) Removing risk reporting
d) Ignoring non-performing assets
Answer: b) Providing early warning signals based on patterns
Explanation: Helps identify stressed accounts early.
Q34. AI-based document processing in banks is mainly used for:
a) Manual handwriting
b) OCR and data extraction from forms & documents
c) Filing in cupboards
d) Printing drafts
Answer: b) OCR and data extraction from forms & documents
Explanation: It automates data entry.
Q35. AI in call centres can:
a) Only play music
b) Analyze calls and provide real-time suggestions to agents
c) Remove recordings
d) Replace telecom switch
Answer: b) Analyze calls and provide real-time suggestions to agents
Explanation: Improves quality and speed of responses.
Q36. AI-based personalization in digital banking apps means:
a) Same offers for all customers
b) Customised product offers based on user behaviour and profile
c) Removing all offers
d) Offline-only products
Answer: b) Customised product offers based on user behaviour and profile
Explanation: AI analyzes behaviour for cross-sell/up-sell.
Q37. Which area is AI MOST used in capital markets?
a) Physical cheque clearing
b) Algorithmic trading and risk analysis
c) Locker allotment
d) Passbook printing
Answer: b) Algorithmic trading and risk analysis
Explanation: AI supports high-speed trading and risk modelling.
Q38. AI helps in KYC by:
a) Ignoring ID documents
b) Automating face match and document verification
c) Removing customer data
d) Printing KYC forms
Answer: b) Automating face match and document verification
Explanation: AI speeds up digital onboarding.
Q39. In internal audit, AI can:
a) Increase manual sampling
b) Analyse 100% transactions and highlight unusual cases
c) Stop all reporting
d) Remove compliance
Answer: b) Analyse 100% transactions and highlight unusual cases
Explanation: AI supports continuous auditing. 👉 (HIGHLY IMPORTANT)
Q40. AI-based chatbots reduce:
a) Access to services
b) Customer satisfaction
c) Response time and operational cost
d) Digital reach
Answer: c) Response time and operational cost
Explanation: Automates routine customer interactions.
CHAPTER 4: RISKS, ETHICS & RECENT DEVELOPMENTS (10 MCQs)
Q41. A key regulatory concern with AI in banking is:
a) Colour of devices
b) Explainability and fairness of AI decisions
c) Number of chairs in office
d) Building paint
Answer: b) Explainability and fairness of AI decisions
Explanation: Regulators want transparent and non-discriminatory models. 👉 (HIGHLY IMPORTANT)
Q42. Data privacy in AI mainly refers to:
a) Sharing customer data freely
b) Protecting personal data and using it lawfully
c) Deleting all data
d) Printing data on notice board
Answer: b) Protecting personal data and using it lawfully
Explanation: Essential for compliance with privacy regulations.
Q43. One of the major RISKS of using AI in lending is:
a) Faster processing time
b) Hidden bias against certain groups in data
c) More documentation errors
d) Reduction in digital services
Answer: b) Hidden bias against certain groups in data
Explanation: Biased models can lead to unfair decisions.
Q44. “Human-in-the-loop” in AI decision systems means:
a) Humans are completely removed
b) Human oversight is involved in key decisions
c) Only machine decides
d) Only manual system is used
Answer: b) Human oversight is involved in key decisions
Explanation: Reduces risk of wrong or unfair fully automated decisions.
Q45. Model governance for AI in banks includes:
a) Only hardware management
b) Policies, validation, monitoring and documentation of models
c) Locker key management
d) Uniform design decisions
Answer: b) Policies, validation, monitoring and documentation of models
Explanation: Ensures safe and controlled use of AI.
Q46. AI systems should be periodically reviewed because:
a) Staff changes
b) Data patterns and business conditions change over time
c) Buildings are renovated
d) Branch codes change frequently
Answer: b) Data patterns and business conditions change over time
Explanation: To manage model drift and keep accuracy high.
Q47. One ethical principle in AI use is:
a) Opacity
b) Accountability
c) Ignoring consent
d) Unlimited data sharing
Answer: b) Accountability
Explanation: Someone must be responsible for AI outcomes.
Q48. An AI system used for credit approval without proper documentation and testing can lead to:
a) Strong compliance
b) Regulatory breaches and customer complaints
c) Better manual control
d) Lower credit volume
Answer: b) Regulatory breaches and customer complaints
Explanation: Regulators require tested and documented models.
Q49. Banks using AI must ensure:
a) Only speed, no fairness
b) Accuracy, fairness, transparency and customer protection
c) Only profit
d) Only marketing campaigns
Answer: b) Accuracy, fairness, transparency and customer protection
Explanation: These are core regulatory expectations. 👉 (HIGHLY IMPORTANT)
Q50. Overall, the impact of AI on banking can be summarized as:
a) Only job loss
b) More efficient, data-driven and personalised banking with new risks to manage
c) No change in services
d) Only increased paperwork
Answer: b) More efficient, data-driven and personalised banking with new risks to manage
Explanation: AI brings major benefits but also new risks needing controls.
🔥 25 Latest & Tricky MCQs on AI (Expected for 2025 Exams)
🔷 Section A: Latest Technology Trends (10 MCQs)
Q1. Generative AI refers to AI systems that:
a) Only classify existing data
b) Create new content such as text, images & audio
c) Store historical data
d) Manage hardware
Answer: b) Create new content such as text, images & audio
Explanation: Generative AI models like ChatGPT create new content. 👉 (HIGHLY IMPORTANT)
Q2. Which technology powers Generative AI models?
a) Linear Regression
b) Deep Learning / Neural Networks
c) Punch card systems
d) Paging algorithms
Answer: b) Deep Learning / Neural Networks
Explanation: GenAI uses advanced deep neural networks like transformers.
Q3. LLM in AI stands for:
a) Large Language Model
b) Loaded Learning Mechanism
c) Long Latency Module
d) Large Loop Memory
Answer: a) Large Language Model
Explanation: LLMs process & generate human-like language. 👉 (HIGHLY IMPORTANT)
Q4. Transformers are mainly used in:
a) Switching electrical circuits
b) NLP and LLM training
c) Database indexing
d) Web routing
Answer: b) NLP and LLM training
Explanation: Transformers learn patterns via attention mechanism.
Q5. Which of the following is an example of Generative AI usage in banking?
a) Printing ATM receipts
b) Automated email drafting and personalized customer communication
c) Cleaning ATM screens
d) Card embossing
Answer: b) Automated email drafting and personalized customer communication
Q6. RLHF in AI training means:
a) Real Learning for Human Finance
b) Reinforcement Learning from Human Feedback
c) Real-life Hardware Function
d) Rapid Language High Framework
Answer: b) Reinforcement Learning from Human Feedback
Explanation: Used in training LLM models like ChatGPT.
Q7. Which emerging concept refers to AI models reasoning like humans?
a) Logical AI
b) AGI (Artificial General Intelligence)
c) Computer Networks
d) OOP
Answer: b) AGI (Artificial General Intelligence)
Explanation: AGI aims for human-like reasoning abilities.
Q8. Federated learning in AI allows:
a) Sharing raw customer data easily
b) Training models on distributed data without central storage
c) Manual collection of data
d) Offline training only
Answer: b) Training models on distributed data without central storage
Explanation: Helps privacy-preserving training. 👉 (HIGHLY IMPORTANT)
Q9. Responsible AI focuses on:
a) Only revenue growth
b) Ethical, safe & fair use of AI models
c) Maximum automation without control
d) Full data visibility to public
Answer: b) Ethical, safe & fair use of AI models
Q10. Quantum AI will mainly improve:
a) ATM printing capacity
b) AI speed & complex problem solving
c) Cheque book distribution
d) Manual approvals
Answer: b) AI speed & complex problem solving
🔷 Section B: Banking-Focused & Regulatory MCQs (10 MCQs)
Q11. RBI’s main focus in AI model governance is:
a) Marketing campaigns
b) Explainability, fairness and transparency
c) Building large offices
d) Extending branch staff
Answer: b) Explainability, fairness and transparency 👉 (HIGHLY IMPORTANT)
Q12. AI-based real-time fraud detection mainly uses:
a) Rule-based static checks
b) Behavioural pattern analysis & anomaly detection
c) Only branch approval
d) Manual ledgers
Answer: b) Behavioural pattern analysis & anomaly detection
Q13. AI in credit underwriting helps banks to:
a) Increase manual paperwork
b) Predict borrower repayment probability using multiple data points
c) Issue loans without assessment
d) Leave decisions to branch manager only
Answer: b) Predict borrower repayment probability using multiple data points
Q14. Use of Generative AI in internal audit enables:
a) Ignoring irregularities
b) Automated summarization & exception highlighting
c) Manual report writing
d) Removing compliance
Answer: b) Automated summarization & exception highlighting
Q15. AI-driven regulatory reporting can:
a) Delay submission timelines
b) Reduce reporting errors and ensure fast analytics
c) Increase paperwork
d) Stop digitization completely
Answer: b) Reduce reporting errors and ensure fast analytics
Q16. Cognitive AI refers to:
a) Manual reasoning
b) Machines mimicking human thinking & decision reasoning
c) Physical robotic arms
d) Only image scanning
Answer: b) Machines mimicking human thinking & decision reasoning
Q17. Model explainability is important because:
a) Helps customers & regulators understand why decisions were made
b) Reduces automation
c) Restricts digital banking
d) Increases cash transactions
Answer: a) Helps customers & regulators understand why decisions were made 👉 (HIGHLY IMPORTANT)
Q18. A key risk of using AI in lending decisions is:
a) Improved approval speed
b) Hidden algorithmic bias
c) Reduced digital access
d) Lower accuracy
Answer: b) Hidden algorithmic bias
Q19. In customer analytics, AI helps banks to:
a) Offer generic services only
b) Personalize offers based on behaviour patterns
c) Block digital channels
d) Remove mobile apps completely
Answer: b) Personalize offers based on behaviour patterns
Q20. Virtual assistants in banking can:
a) Replace ATMs
b) Provide instant customer support & reduce service costs
c) Replace database storage
d) Print cheques
Answer: b) Provide instant customer support & reduce service costs
🔷 Section C: Expected Questions for 2025 (5 MCQs)
Q21. What will be the most critical regulatory expectation for AI-based lending in 2025?
a) Data secrecy & explainable decisioning
b) Unlimited automation
c) Zero documentation
d) Removal of human involvement
Answer: a) Data secrecy & explainable decisioning 👉 (HIGHLY IMPORTANT)
Q22. What is the fastest-growing use case of Generative AI in banks?
a) Manual cheque verification
b) Automated document summarization & report drafting
c) Fixing ATMs
d) Counting cash
Answer: b) Automated document summarization & report drafting
Q23. Which new AI trend will help banks handle highly unstructured data?
a) Rule-based systems
b) Foundation models and LLMs
c) Tele-banking only
d) Fax processing
Answer: b) Foundation models and LLMs
Q24. AI will help RBI in supervision by:
a) Manual inspections only
b) Real-time monitoring of large data & risk indicators
c) Removing financial oversight
d) Reducing regulation
Answer: b) Real-time monitoring of large data & risk indicators
Q25. The biggest opportunity area for AI in Indian banking in 2025-26 is:
a) AI-based fraud monitoring & compliance automation
b) Passbook printing
c) Building more manual counters
d) Offline customer enrolment
Answer: a) AI-based fraud monitoring & compliance automation 👉 (HIGHLY IMPORTANT)