Wayground logo

Free Printable Worksheets

Font size

S
M
L
XL
Worksheets

Module 1 – Core Concepts of AI

Total questions: 106

Worksheet time: 53mins

Name
Class
Date
1.

What best describes artificial intelligence?

a)

A single software product

b)

A branch of computer science focused on simulating intelligent behaviour

c)

Any automated workflow

d)

A deterministic decision system

2.

Which element is commonly found in definitions of AI?

a)

Forecasting

b)

Intelligence

c)

Compliance

d)

Profit optimisation

3.

Which characteristic most differentiates AI from traditional automation?

a)

Speed

b)

Accuracy

c)

Learning and adaptation

d)

Cost efficiency

4.

AI systems are typically designed to be:

a)

Random

b)

Goal-directed

c)

Fully transparent

d)

Deterministic

5.

Why has the definition of AI evolved over time?

a)

Laws have changed

b)

Computing costs decreased

c)

Societal expectations of intelligence have changed

d)

AI no longer uses data

6.

Which output is commonly generated by AI systems?

a)

Legal opinions

b)

Predictions

c)

Policies

d)

Guarantees

7.

Which feature of AI increases governance complexity?

a)

Low cost

b)

Opacity

c)

Documentation

d)

Standardisation

8.

What does AI autonomy refer to?

a)

AI operating without electricity

b)

AI making decisions without direct human oversight

c)

AI replacing humans entirely

d)

AI operating legally

9.

Why does speed and scale matter for AI governance?

a)

It reduces system costs

b)

It limits deployment

c)

Errors can propagate rapidly

d)

It improves explainability

10.

What does probabilistic output mean?

a)

Outputs are always random

b)

Outputs are based on likelihood, not certainty

c)

Outputs are unverifiable

d)

Outputs are human-approved

11.

Which category of AI currently exists in real-world use?

a)

AGI

b)

ASI

c)

ANI

d)

Sentient AI

12.

AGI is best described as:

a)

AI that performs one task well

b)

AI with human-level general intelligence

c)

AI that operates autonomously

d)

AI used in healthcare

13.

Why are AGI and ASI important to governance discussions?

a)

They are widely deployed

b)

They define current compliance rules

c)

They inform future risk planning

d)

They replace ANI

14.

Broad AI is best described as:

a)

Fully human-like AI

b)

A single narrow model

c)

Multiple AI systems working together

d)

Non-learning automation

15.

AI agents are an example of:

a)

ANI

b)

Broad AI

c)

AGI

d)

ASI

16.

Why do autonomous vehicles qualify as broad AI?

a)

They use sensors

b)

They combine multiple AI systems

c)

They are fully sentient

d)

They eliminate human oversight

17.

What is the purpose of the OECD AI classification framework?

a)

Certify AI vendors

b)

Replace national laws

c)

Support common understanding and risk assessment

d)

Rank AI performance

18.

Which is NOT a goal of the OECD framework?

a)

Inform AI inventories

b)

Support sector-specific frameworks

c)

Guarantee AI safety

d)

Promote common understanding

19.

Which OECD dimension relates to how AI decisions affect society?

a)

AI model

b)

Tasks and output

c)

Economic context

d)

Data input

20.

Which dimension covers what the AI system does?

a)

Data and input

b)

Tasks and output

c)

People and planet

d)

Economic context

21.

Why is data dependency a governance concern?

a)

Data is expensive

b)

Data quality affects outcomes and bias

c)

Data is always personal

d)

Data eliminates risk

22.

Which megatrend most directly enables scalable AI deployment?

a)

Blockchain

b)

Cloud computing

c)

Social media

d)

Robotics

23.

How does IoT interact with AI?

a)

It limits data collection

b)

It provides large volumes of real-time data

c)

It replaces machine learning

d)

It reduces governance needs

24.

Why do mobile technology and social media matter for AI?

a)

They reduce bias

b)

They generate large behavioural datasets

c)

They replace cloud computing

d)

They ensure compliance

25.

Privacy-enhancing technologies are relevant because they:

a)

Remove the need for AI governance

b)

Enable AI use while reducing privacy risks

c)

Replace consent

d)

Eliminate data use

26.

Blockchain’s relationship to AI is best described as:

a)

Competitive

b)

Unrelated

c)

Complementary in some use cases

d)

Required

27.

Why do autonomous weapons raise governance concerns?

a)

Low accuracy

b)

Limited scale

c)

Potential for severe harm

d)

Lack of data

28.

Computer vision is commonly used for:

a)

Legal reasoning

b)

Image and pattern recognition

c)

Policy drafting

d)

Encryption

29.

One benefit of AI in healthcare is:

a)

Eliminating doctors

b)

Faster and more accurate scan analysis

c)

Guaranteed diagnoses

d)

Removing liability

30.

AI’s similarity to big data lies in:

a)

Storage requirements

b)

Volume, velocity, and variety of data processing

c)

Legal frameworks

d)

Manual review

31.

Why can AI reduce human error?

a)

AI is unbiased

b)

AI automates repetitive tasks

c)

AI is deterministic

d)

AI replaces judgement

32.

Why can AI also introduce bias?

a)

AI lacks goals

b)

AI depends on training data

c)

AI is autonomous

d)

AI is opaque

33.

Recognition use cases include:

a)

Route optimisation

b)

Image or speech recognition

c)

Forecasting demand

d)

Risk scoring

34.

Forecasting use cases include:

a)

Facial recognition

b)

Predicting demand or trends

c)

Chatbots

d)

Optimisation

35.

Personalisation use cases aim to:

a)

Standardise outputs

b)

Tailor experiences to individuals

c)

Remove humans

d)

Reduce data use

36.

Goal-driven optimisation focuses on:

a)

Detecting fraud

b)

Maximising outcomes under constraints

c)

Interpreting text

d)

Ensuring fairness

37.

Recommendation systems are an example of:

a)

Deterministic automation

b)

Goal-driven optimisation

c)

Static rules

d)

Manual decision-making

38.

Why is human interaction important in AI?

a)

AI cannot operate alone

b)

Humans shape outcomes and impacts

c)

Humans approve every output

d)

Humans reduce accuracy

39.

Which challenge is central to AI governance?

a)

Eliminating innovation

b)

Balancing innovation and risk management

c)

Avoiding all automation

d)

Reducing computation

40.

Why is understanding AI characteristics critical?

a)

To improve marketing

b)

To manage inherent risks

c)

To maximise profit

d)

To replace compliance

41.

Which characteristic increases misuse risk?

a)

Transparency

b)

Autonomy

c)

Documentation

d)

Testing

42.

What makes AI governance different from traditional IT governance?

a)

Lower costs

b)

Probabilistic outputs and learning behaviour

c)

Shorter lifecycles

d)

Static behaviour

43.

AI governance must consider:

a)

Only technology

b)

Only data

c)

Technical and societal impacts

d)

Only legal compliance

44.

Why is opacity problematic?

a)

It slows systems

b)

It limits explainability and accountability

c)

It increases cost

d)

It prevents deployment

45.

What role does learning play in AI risk?

a)

It removes risk

b)

It can change behaviour post-deployment

c)

It ensures fairness

d)

It simplifies monitoring

46.

Which statement is true about AI and humans?

a)

AI replaces all human judgement

b)

AI outcomes are shaped by human choices

c)

AI eliminates governance

d)

AI is value-neutral

47.

Which AI characteristic most challenges predictability?

a)

Speed

b)

Probabilistic behaviour

c)

Data storage

d)

Hardware

48.

Why must AI governance be lifecycle-based?

a)

AI is expensive

b)

AI behaviour can change over time

c)

AI is always autonomous

d)

AI is fully transparent

49.

Which best describes AI’s societal impact?

a)

Fixed and predictable

b)

Independent of humans

c)

Influenced by deployment and use

d)

Limited to technology teams

50.

What is the key takeaway of Module 1?

a)

AI is only a technical issue

b)

AI requires governance due to unique characteristics and risks

c)

AI eliminates human decision-making

d)

AI governance replaces innovation

51.

What best defines an AI model?

a)

A dataset used to train algorithms

b)

A program that applies algorithms to data to make predictions or decisions

c)

A computing infrastructure

d)

A user interface

52.

What does the term algorithm refer to?

a)

A trained AI system

b)

A set of instructions or rules to solve a problem

c)

A data repository

d)

A computing environment

53.

Which term describes the full operational environment including models, data, and infrastructure?

a)

Model

b)

Algorithm

c)

Dataset

d)

System

54.

Machine learning differs from traditional programming because it:

a)

Uses fixed rules

b)

Learns patterns from data

c)

Requires no data

d)

Is deterministic

55.

Which relationship between AI categories is correct?

a)

AI⊂ML⊂DL⊂GenAI\text{AI} \subset \text{ML} \subset \text{DL} \subset \text{GenAI}

b)

ML⊂AI⊂DL⊂GenAI\text{ML} \subset \text{AI} \subset \text{DL} \subset \text{GenAI}

c)

GenAI⊂DL⊂ML⊂AI\text{GenAI} \subset \text{DL} \subset \text{ML} \subset \text{AI}

d)

DL⊂GenAI⊂ML⊂AI\text{DL} \subset \text{GenAI} \subset \text{ML} \subset \text{AI}

56.

What is a defining characteristic of deep learning?

a)

Rule-based inference

b)

Multi-layer neural networks

c)

Small datasets

d)

Deterministic outputs

57.

One advantage of deep learning over traditional ML is:

a)

Lower data requirements

b)

Manual feature extraction

c)

Automatic feature learning

58.

A key limitation of deep learning models is that they:

a)

Cannot process images

b)

Require large amounts of data and compute

c)

Do not scale

d)

Are rule-based

59.

Generative AI systems are designed to:

a)

Optimise numerical values

b)

Classify existing data

c)

Generate new content

d)

Apply business rules

60.

Which is a common ethical risk of generative AI?

a)

Overfitting

b)

Deterministic bias

c)

Misinformation

d)

Feature scaling

61.

Agentic AI systems are characterised by:

a)

Static outputs

b)

Autonomous decision-making and action

c)

Human-in-the-loop control only

d)

Rule-based logic

62.

Which learning approach uses labelled data?

a)

Unsupervised learning

b)

Reinforcement learning

c)

Supervised learning

d)

Agentic learning

63.

A challenge of supervised learning is:

a)

Low accuracy

b)

Lack of scalability

c)

Need for labelled data

d)

No structure

64.

Unsupervised learning is best suited for:

a)

Regression tasks

b)

Customer segmentation

c)

Policy enforcement

d)

Classification with labels

65.

Reinforcement learning differs from supervised learning because it:

a)

Uses labelled data

b)

Learns through rewards and penalties

c)

Is deterministic

d)

Requires no environment

66.

Which algorithm is commonly used for numeric prediction?

a)

Logistic regression

b)

Linear regression

c)

Decision trees

d)

CNNs

67.

Logistic regression is primarily used for:

a)

Continuous prediction

b)

Binary classification

c)

Image recognition

d)

Text generation

68.

Random forests improve performance by:

a)

Using a single tree

b)

Combining multiple decision trees

c)

Eliminating bias

d)

Reducing training data

69.

Neural networks are especially useful for:

a)

Simple rules

b)

Complex pattern recognition

c)

Data storage

d)

Compliance checks

70.

Transformer models are important because they:

a)

Process data sequentially only

b)

Capture contextual relationships efficiently

c)

Require labelled datasets only

d)

Are deterministic

71.

Multimodal models differ from language models because they:

a)

Process only text

b)

Handle multiple data types

c)

Are smaller

d)

Eliminate bias

72.

Which governance concern is heightened by multimodal models?

a)

Performance tuning

b)

Privacy risk

c)

Cost reduction

d)

Deployment speed

73.

Retrieval-augmented generation improves GenAI by:

a)

Reducing model size

b)

Incorporating external information

c)

Eliminating hallucinations completely

d)

Replacing training

74.

Proprietary models typically:

a)

Are fully transparent

b)

Are controlled by vendors

c)

Require no governance

d)

Cannot be deployed

75.

Open-source models raise governance concerns because:

a)

They are illegal

b)

Accountability may be unclear

c)

They do not scale

d)

They cannot be secured

76.

Large language models differ from small ones mainly in:

a)

Existence

b)

Parameter count and resources

c)

Output type

d)

Legal status

77.

Small language models are often preferred when:

a)

Broad versatility is required

b)

Resources are constrained

c)

Multimodality is required

d)

Data is unlimited

78.

Language models are best suited for:

a)

Image recognition

b)

Text-based tasks

c)

Sensor fusion

d)

Video generation

79.

Why might organisations combine multiple model types?

a)

To avoid governance

b)

To handle complex tasks

c)

To reduce costs only

d)

To eliminate humans

80.

Expert systems differ from ML models because they:

a)

Learn autonomously

b)

Use rule-based inference

c)

Require large datasets

d)

Are probabilistic

81.

The inference engine in an expert system:

a)

Stores data

b)

Applies rules to reach conclusions

c)

Collects sensor inputs

d)

Trains models

82.

Why are expert systems more explainable?

a)

They are smaller

b)

They are rule-based

c)

They are open source

d)

They are generative

83.

Which model type raises the greatest opacity concerns?

a)

Linear regression

b)

Decision trees

c)

Deep neural networks

d)

Expert systems

84.

Why do large models increase bias risk?

a)

They are deterministic

b)

They rely on massive datasets

c)

They use rules

d)

They are transparent

85.

Which learning method involves exploration vs exploitation?

a)

Supervised

b)

Unsupervised

c)

Reinforcement

d)

Semi-supervised

86.

Why is reinforcement learning risky in high-impact domains?

a)

It lacks labels

b)

It may learn harmful behaviours

c)

It cannot scale

d)

It is deterministic

87.

Which architecture is best for image recognition?

a)

Linear regression

b)

CNN

c)

Logistic regression

d)

RNN

88.

Graph neural networks are best suited for:

a)

Text generation

b)

Relationship-based data

c)

Image processing

d)

Regression

89.

Why does agentic AI raise governance concerns?

a)

Low accuracy

b)

Autonomous actions

c)

Limited data

d)

Deterministic logic

90.

Which is a key governance risk of proprietary GenAI?

a)

Low innovation

b)

Limited transparency

c)

Poor performance

d)

No data

91.

Multimodal models raise additional privacy risks because they:

a)

Are smaller

b)

Combine multiple data types

c)

Are open source

d)

Are deterministic

92.

Why should governance professionals understand ML basics?

a)

To build models

b)

To audit source code

c)

To assess risk and engage technical teams

d)

To replace engineers

93.

Key takeaway of this module is that:

a)

All models are equivalent

b)

Model types determine capabilities and risks

c)

Governance is purely technical

d)

Larger models are always better

94.

Which training method is most appropriate when outcomes are unknown and patterns must be discovered?

a)

Supervised learning

b)

Reinforcement learning

c)

Unsupervised learning

d)

Deterministic programming

95.

Semi-supervised learning is valuable because it:

a)

Eliminates bias

b)

Requires only labelled data

c)

Reduces the need for extensive manual labelling

d)

Produces deterministic results

96.

Which model type is most likely to require the greatest computational resources?

a)

Linear regression

b)

Decision trees

c)

Large language models

d)

Expert systems

97.

Why are foundation models important in modern AI systems?

a)

They replace governance

b)

They enable reuse across many tasks

c)

They eliminate training data

d)

They are deterministic

98.

Which AI architecture is best suited for sequential data such as text or speech?

a)

CNN

b)

RNN

c)

Linear regression

d)

Decision trees

99.

From a governance perspective, why does model size matter?

a)

It affects branding

b)

It influences risk, cost, and bias exposure

c)

It guarantees accuracy

d)

It eliminates oversight

100.

Which statement best reflects responsible AI model selection?

a)

Always choose the most advanced model

b)

Select models aligned to purpose, risk, and constraints

c)

Prefer open source in all cases

d)

Avoid multimodal models

101.

What best defines an AI model?

a)

A dataset used to train algorithms

b)

A program that applies algorithms to data to make predictions or decisions

c)

A computing infrastructure

d)

A user interface

102.

What does the term “algorithm” refer to?

a)

A trained AI system

b)

A set of instructions or rules to solve a problem

c)

A data repository

d)

A computing environment

103.

Which term refers to the full operational environment of AI?

a)

Model

b)

Algorithm

c)

Dataset

d)

System

104.

Machine learning is best described as:

a)

Explicitly programmed decision-making

b)

Learning patterns from data without explicit programming

c)

Manual feature engineering

d)

Deterministic automation

105.

Which category relationship is correct?

a)

AI ⊂ ML ⊂ DL ⊂ GenAI

b)

ML ⊂ AI ⊂ DL ⊂ GenAI

c)

GenAI ⊂ DL ⊂ ML ⊂ AI

d)

DL ⊂ GenAI ⊂ ML ⊂ AI

106.

What distinguishes deep learning from traditional ML?

a)

Deterministic outputs

b)

Multi-layered neural networks

c)

Rule-based logic

d)

Manual feature engineering