ISC Class 12 Artificial Intelligence Specimen Paper 2026-27: Download PDF for CISCE Board Exam

Category

Details

Conducting Body

Council for the Indian School Certificate Examinations (CISCE)

Exam Name

ISC Class 12 Artificial Intelligence

Academic Session

2026–27

Exam Date

To be announced 

Paper Type

Theory

Total Marks

70 Marks (Internal Assessment: 30 Marks)

Exam Duration 

3 Hours

Official Website

cisce.org

SECTION A – 16 MARKS

Question 1

(i) Which of the following is an example of sentiment analysis? (Understand) [1]
(a) Translating reviews from Hindi to English
(b) Grouping similar reviews into clusters without labels
(c) Extracting names of products and brands from reviews
(d) Detecting whether a review is positive, negative, or neutral

(ii) Choose the statements that justify the preference for HSV / HSL colour model over RGB for colour detection.
(P) Lightness and value are not considered.
(Q) Saturation is separated from brightness.
(R) Hue is separated from brightness.
(a) Only (P) and (Q)
(b) Only (Q) and (R)
(c) Only (P) and (R)
(d) (P), (Q) and (R)

(iii) The learning paradigm which involves an AI system, learning to maximise long-term rewards through a sequence of decisions is:
(a) Deep learning
(b) Supervised learning
(c) Unsupervised learning
(d) Reinforcement learning

(iv) Given below are two statements marked Assertion and Reason. Read the statements carefully and choose the correct option.
Assertion: The ReLU activation function is preferred to sigmoid in the hidden layers of a deep neural network.
Reason: For large positive or negative inputs, the sigmoid function produces near zero derivatives, causing weight updates in early layers to become negligibly small, a problem known as the vanishing gradient.
(a) Both Assertion and Reason are true, and Reason is the correct explanation for the Assertion.
(b) Both Assertion and Reason are true, but Reason is not the correct explanation for the Assertion.
(c) Assertion is true and Reason is false.
(d) Both Assertion and Reason are false.

(v) A face-recognition system at an airport achieves 99% overall accuracy but only 70% accuracy for passengers with different skin tones. Which one of the following is MOST likely the cause for this bias?
(a) Insufficient computing power at the airport
(b) Faulty algorithm that cannot model non-linear facial features
(c) Under-represented training data, embedding demographic bias
(d) Overfitting to the training data across all demographic groups

(vi) Given below are two statements marked Assertion and Reason. Read the statements carefully and choose the correct option.
Assertion: Medical diagnosis systems based on Machine Learning can improve the efficiency of healthcare services.
Reason: Genetic information is one of the sources of data that can be analysed by Machine Learning models.
(a) Both Assertion and Reason are true, and Reason is the correct explanation for the Assertion.
(b) Both Assertion and Reason are true, but Reason is not the correct explanation for the Assertion.
(c) Assertion is true and Reason is false.
(d) Both Assertion and Reason are false.

(vii) Which of the following Neural Network (NN) architectures is suited for image classification tasks?

(viii) A neural network used for recognising handwritten digits contains an input layer, one or more hidden layers and an output layer. The hidden layers transform the input information before the final prediction is made. What is the primary role of the hidden layers in such a neural network?
(a) To display the final predicted class
(b) To divide the dataset into training and testing sets
(c) To store the original input data without modification
(d) To learn and transform intermediate features from the input data

(ix) A bank uses a Machine Learning system to identify unusual transaction patterns that may indicate identity theft. Which practical implication of Machine Learning is evident here?
(a) Fraud Detection
(b) Medical Diagnosis
(c) Workflow automation
(d) Predictive Maintenance

(x) The Computer Vision technique that can be used to count the number of people in a photograph is:
(a) Object tracking
(b) Object detection
(c) Facial recognition
(d) Image classification

(xi) Name the Deep Learning model commonly used for image classification.

(xii) A machine learning model incorrectly classifies a genuine credit card transaction as fraudulent. Name the type of prediction made by the model.

(xiii) What is the purpose of the fit() function while training an Artificial Neural Network in Keras?

(xiv) A hospital uses Machine Learning to analyse patient records for disease prediction. Why should the model be updated periodically with new patient data?

🡇

🡇

🡇

To download complete Question Paper PDF along with marking scheme for ICSE Class 10 physics Specimen Paper 2027 for board exams. Students can tap on the link below and get the PDF for free to download and start their preparation in advance to score good grades in 2027 board exams. 

DOWNLOAD: ICSE Class 12 Artificial Intelligence Specimen Paper 2027 PDF