CAIIB · Advanced Business & Financial Management · Module D — Emerging Business Solutions
CAIIB ABFM Unit 22 — Artificial Intelligence: Free MCQs with Explanations
By Swarnakshi Jha (ex-banker) · Updated 29 October 2026 · Paper III of the December 2026 cycle (Sunday 13 December 2026)
This is the free practice page for CAIIB ABFM Unit 22 — Artificial Intelligence (Module D — Emerging Business Solutions). The CAIIB Mock Centre holds 180 questions on this unit alone, every one with a why-right / why-wrong explanation; below are the facts examiners keep returning to, five practice questions with explanations, and the traps that cost marks.
ABFM · Module D180 MCQs in the Mock Centre5 free belowPaper on 13 December
Key facts examiners test from this unit
AI: meaning, history and applications — AI makes computers learn and solve problems like people; Turing gave the test and McCarthy the name (1956).
Google's contributions to AI — RankBrain reads meaning, the AI Hub shares ML tools, Kubeflow runs pipelines, Duplex makes calls.
AI in banking and finance, and its future scope — banks use AI to serve, advise, predict, protect and lend; hybrid systems lead in financial prediction.
Changed since the textbook
India AI Governance Guidelines (5 November 2025). MeitY released the India AI Governance Guidelines under the IndiaAI Mission. They adopt seven principles that mirror the FREE-AI Sutras, prefer focused amendments to existing laws over a separate AI law, and propose an AI Governance Group, an AI Safety Institute, graded liability by role and risk, and a national database of AI incidents. Source: AZB & Partners
RBI's FREE-AI framework for the financial sector (13 August 2025). The RBI released the report of its Committee on the Framework for Responsible and Ethical Enablement of AI (FREE-AI), set up on 26 December 2024 under Dr Pushpak Bhattacharyya of IIT Bombay. Source: RBI press release
5 free practice questions — tap an option
Five practice questions from this unit. Answers lock on the first tap, exactly as in the exam, and an explanation opens underneath. The Mock Centre has all 180 questions on this unit, from easy to tough, with caselets and key-in numericals, why every other option is wrong and the exam tip — and tracks your score and revision dates.
Question 1 of 5
Besides making the Assistant identify itself, what other changes were made to Duplex after the controversy?
Why (a): After the controversy, the Assistant must identify itself and warn call recipients that the conversation is recorded, and businesses may choose not to take part in Duplex. (b) is wrong: The change was to warn recipients that calls are recorded, not to delete recordings. (c) is wrong: Businesses may opt out; they need not opt in first. (d) is wrong: The textbook does not describe a human listening in; the safeguards are disclosure and opt-out.
Question 2 of 5
How does the textbook describe the range of problems neural networks can model?
Why (c): The textbook calls neural networks highly adaptable algorithms for a wide variety of modelling problems, supervised and unsupervised. (a) is wrong: They model business data such as credit and fraud records too. (b) is wrong: The textbook says they suit unsupervised problems as well. (d) is wrong: They do best with many inputs that have non-linear links to the target.
Question 3 of 5
Why is knowing a sentence's structure not enough for a computer to understand it, according to the textbook?
Why (a): The textbook says understanding language needs knowledge of the subject and the context as well as the structure of sentences; this is why knowledge representation, putting knowledge in a form a computer can reason with, matters. (b) is wrong: The textbook does not say grammar is too large to store; Chomsky's theory was formal enough to be programmed. (c) is wrong: The textbook does not say structure varies by speaker; the gap is that meaning needs more than structure. (d) is wrong: Chomsky's syntax was formal enough to be programmed, so machines can handle structure.
Question 4 of 5
What role do upper ontologies play in knowledge representation?
Why (c): Upper ontologies are the most general: they define broad concepts such as object, event and time, providing a base for all other knowledge and mediating between domain ontologies, which each cover one field. (a) is wrong: Covering one field in depth describes a domain ontology. (b) is wrong: Concepts for one application form an application ontology, not an upper one. (d) is wrong: Facts about individual customers are instance data, not an ontology.
Question 5 of 5
Which kind of logic uses quantifiers and predicates to express facts about objects, their properties and relations?
Why (d): First-order logic includes quantifiers and predicates in its syntax to express facts about objects, their qualities and their relationships. (a) is wrong: Modal logics are extensions for particular domains. (b) is wrong: Propositional logic rests on truth functions such as 'or' and 'not'. (c) is wrong: Fuzzy logic deals with degrees of truth.
0 of 5 answered.
Traps in this unit
McCarthy named AI; Turing proposed the test.
Machine learning is part of AI, not separate from it.
Robo-advice is debated, not settled.
Practise all 180 questions on Unit 22. The free CAIIB demo opens Units 1–2 of every paper with full explanations; the ABFM Online Mock Centre unlocks all 4,000+ ABFM questions, the full-length papers with caselets and key-ins, module mocks and Smart Revision for 18 months (₹1,499).
Independent study aid — not affiliated with, endorsed by, or sponsored by the Indian Institute of Banking & Finance (IIBF). Exam pattern, dates, fees and passing criteria are set by IIBF — always confirm the current rules at iibf.org.in. CAIIB content follows rules and developments up to IIBF's cut-off of 30 June 2026 (where a rule changed later, the update box says so); verify from official sources before relying on them in practice.