Answer Key

Artificial Intelligence - 9
Part B - Subject-specific Skills: Artificial Intelligence
Unit 1: Introduction to Artificial Intelligence

Let’s Review (Pg. 127)

A. Match the following.
1. – d
2. – c
3. – e
4. – b
5. – a
B. (To be attempted by the students)

AI Review (Pg. 140)

Identify the SDG by the target given below.
1. SDG-16 2. SDG-14 3. SDG-6
4. SDG-5 5. SDG-1

Exercises (Pgs. 155-157)

A. Tick (✓) the correct answers.
1. b. 2. c. 3. b. 4. b. 5. b.
6. a. 7. c. 8. b. 9. c. 10. c.
B. Fill in the blanks.
1. IBM – Deep Blue 2. Turing test 3. 17
4. 2030 5. No Poverty 6. Roomba
7. NLG and NLU 8. healthcare 9. machine learning
10. Artificial Superintelligence
C. Write T for True and F for False.
1. True 2. True 3. False 4. True 5. True
6. False 7. True 8. True 9. True 10. True
D. Give one word answers for the following.
1. Ethics 2. Good health and well-being
3. Artificial Narrow Intelligence (ANI) 4. Artificial Superintelligence (ASI)
5. Alan Turing 6. Smart home
7. Natural Language Processor 8. Rio de Janerio
9. Machine Learning Engineer 10. Speech Recognition
E. Long answer type questions
1. Artificial intelligence is a technology, a form of intelligence, and a field of study that emphasises
the development of intelligent systems (both hardware and software) that are capable of
perceiving their environment, learning from it, thinking, and reasoning. They have capabilities like
humans, such as speech recognition, learning, planning, decision-making and problem-solving.
Many automotive companies are using AI to enable car owners ‛to talk to their cars’. Tesla Bot is
one such example. AI is also used in self-driven cars to identify obstacles in the path and to take
decisions while driving to make the journey safe.
Similarly, AI is used in healthcare industry to diagnose diseases and assist doctors in making
decisions. One such example is IBM’s Watson.
2. Artificial Intelligence can be categorised into various types. There are mainly two types of
classifications, which are based on functionalities and capabilities of AI.
³³ Types of AI: Based on Functionalities
Based on the functionality of AI-based applications, Artificial Intelligence can be mainly
categorized into four types:
XX Reactive Machines: This was the first form of AI that enabled machines to emulate human
mind to respond to different stimuli. These machines can automatically respond to a
limited combination of inputs. They do not have memory-based learning ability, that is,
they do not require past experiences to react.
XX Limited Memory: Limited Memory machines have learning capabilities and use historical
data to make decisions. Chatbots and virtual assistants are all driven by Limited Memory
Technology.
XX Theory of Mind: Currently, researchers are trying to develop the next level of AI systems
called Theory of Mind. Such systems will be able to understand the need of human beings,
emotions, beliefs, and thought process.
³³ Types of AI: Based on Capabilities
Based on capabilities and learning stage, artificial intelligence is broadly classified into three types.
XX Artificial Narrow Intelligence: Also, known as Weak AI, it is the only type of artificial
intelligence that has been successfully released to date. It has a narrow range of abilities
and can successfully perform singular tasks, such as facial recognition, speech recognition.
Though these machines appear intelligent, but they operate under a narrow set of
constraints.
XX Artificial General Intelligence: It is also referred to as strong AI or Deep AI. It is on par with
human capabilities, such as the ability to think, understand, learn, and apply intelligence
to solve any problem. AGI can think, understand, and perform tasks more efficiently than
humans only for a particular function assigned to them.
XX Artificial Superintelligence: It is a futuristic notion that does not just mimic human
intelligence but supersedes human intelligence. For artificial intelligence to become a
reality, computing programs have to be superior than human intelligence in all parameters.
3. Natural Language Processing (NLP) is the domain of AI that focuses on how to program
computers to process and analyse natural languages, such as English, Hindi, French, and so on.
Natural language means the way we humans communicate with each other, both written and
spoken. The objective behind this feature is to make the interaction between machines and
humans coherent, so that computers can give logical responses to resolve a query. Similar to
humans having a brain to process the input using its organs, computers are attuned to process
the input received. It converts the input into a code understandable for computers. NLP focuses
 on both verbal and written sections of human language. It has two main components—Natural
Language Processing (NLP) and Natural Language Generation (NLG).
³³ Natural Language Generation (NLG): It processes and decodes the sentences and words that
humans use to communicate. Natural Language Generation process involves: Text Planning,
Sentence Planning, and Text Realisation.
³³ Natural Language Understanding (NLU): A sub-field of natural language processing, it involves
understanding the language of written text or pixels and translating it into a language that the
machine understands.
4. Computer vision is an integral part of AI and helps the computer to automatically recognize,
analyse, and interpret the visual data from real-world images in the same way as humans do.
CV not only displays the same efficiency as the human eye but also attempts to surpass human
capability. Computer vision possesses the unique ability of pattern recognition and image
identification, and it uses visuals extracted from various sources, such as digital content, graphics,
documents, and so on to help the system to identify, examine, decode, and choose in real-time
situations. This technology is widely used in various sectors, such as healthcare and defense.
Doctors use MRI scans for diagnosis purposes, due to projected visuals from various angles and
pixels. The military uses drones for security purposes.
5. Face Recognition: Facial recognition and authentication is a vital security application where
computer vision can detect any face and match it with a database of persons under a warrant.
Self-driving cars: Computer vision is used to detect and classify obstacles, such as road signs or
traffic lights, and create 3D maps or motion estimation.
Agriculture: Computer vision is being used to monitor the health of livestock, monitor remote
farms, continuously monitor plant growth, and detect diseases in plants to name a few.
6. A smart city is a city that harnesses Information and Communication Technology (ICT) and the
Internet of Things (IoT) with other advanced innovations to improve the quality of life of its
people. Simply put, a smart city is a concept that makes traditional services more efficient and
sustainable through the use of information and communication technologies to enhance the
quality of life of the inhabitants. Cities are evolving at a faster rate. The population of the world
has increased significantly in the past years and so has the people’s standard of living. Cities, at
present, consume around 75 percent of the world’s resources, which is one of the main causes
of greenhouse gases emission. This can have a negative impact on the environment for years
to come. Hence, this makes the development of smart cities necessary. The aim of developing
smart cities is to integrate technological innovations in fields, such as artificial intelligence and
IoT to deal with infrastructure, social, economic, and ecological challenges, in order to have more
healthy and comfortable life while taking care of our environment.
7. Al is one of the tools to deliver SDGs. Al capabilities are already being used in various ways to
further societal goals. With advances in deep learning, it is easy to tap the self-learning power
of Al to understand and solve complex and interconnected sustainable development goals. By
making smart systems that are intelligent with self-learning capability to understand various
processes, AI can be a big help in achieving SDGs. For example, an AI system developed by a
team of computer scientists and satellite experts at Stanford University is being used to identify
poverty hotspots from satellite images. It is a system of a self-updating world map that scans the
day and night time satellite images of an area and checks for the signs of poverty, such as the
intensity of nightlight, and urban areas.
8. A research scientist is responsible for analysing, designing, and undertaking information.
Expertise in multiple AI disciplines is required. The core responsibilities of a research scientist are
researching, developing, and implementing computer vision, NLP, machine learning, and artificial
intelligence techniques.
 9. Al ethics is concerned with how to minimize the ethical harms that can arise from AI in society
either from inappropriate use of the application or poor design of the system. Al system, if
and when developed with malicious intent or trained with adversarial data, makes it prone to
errors that is why AI ethics is important. The ethical concerns associated with adopting AI are
as follows:
³³ Accountability: When a machine is partially or completely responsible for making a decision,
this forms a new set of ethical concerns. We, as a society need accountability structures when
AI plays a role in decision-making.
³³ AI Bias and Discrimination: An AI system must be designed to minimise bias and
discrimination. AI bias is an anomaly in the results produced by the system. Human beings are
inherently susceptible to biases and are responsible for building AI systems; thus, there are
chances for human bias to be embedded in the systems created by them.
³³ Singularity: It is the hypothetical point at which artificial intelligent machines will surpass
human intelligence and human beings will no longer be the most intelligent beings on earth.
³³ Security: As a result of the design, development, and deployment of an AI system, it can pose
threat to privacy. AI and data are inseparable; the development of AI systems requires a large
amount of data. Thus, there may be several implications of how data is stored, processed,
managed, and used.
³³ Inequality: There are fears that AI technologies are increasing inequality worldwide. The
advances in the AI technologies might increase the risk of monopolisation as the company
having the right data will win.
³³ Job Loss: Automation and AI always engender fear and concern. Some believe that it will
bring mass-scale unemployment and social unrest.
³³ Artificial Stupidity: AI-powered machines are not immune to making mistakes and who should
be held responsible when these machines make mistakes? For example, if a robot does wrong
diagnosis of an illness, who should be held accountable?
10. Artificial Intelligence (AI) has had a significant impact on the education sector. It has
revolutionised the way institutions and teachers work and students learn. AI-enabled tools are
used for developing skills and testing systems. With major advances being made in AI every day,
it is expected that AI can help bridge gaps in learning and teaching. In near future, it is hoped
AI will drive efficiency, personalisation and help streamline admin tasks. By leveraging the best
attributes of machines and teachers, the vision for AI in education is where the two would work
together for maximising the outcome of students.
Artificial intelligence tools have helped overcome language barrier and obstacle posed by
disability (visual or hearing impairments) and made global classrooms available to one and all.
Presentation Translator is a free plug-in for PowerPoint that creates subtitles in real time for
what the teacher is saying. This has thrown open possibilities for students who are unable to
attend school due to illness or who require learning at a different level or on a particular subject
that isn’t available in their own school.
F. Application-based questions
1. AI-based recommendation engines work as data-filtering tools. They make use of various
algorithms and data to recommend the most relevant items to customers. To do so, they extract
data based on customer behaviour, like searches, clicks, and purchases. Then, they analyse and
find out what will appeal most to that particular customer in future. This enables customers to
find products they want to buy.
2. The algorithm analyses words, the occurrence of words, and the distribution of words and phrases
inside the content of the emails and segregates them into spam and non-spam categories.