M.A.

M COLLEGE OF ENGINEERING

TRICHY

Department of Artificial intelligence and Data science

AD8013 - Ethics of Artificial intelligence

Unit-1 Introduction to ethics of Ai

Artificial Intelligence (AI) and its multiple sub-domains are being

increasingly employed in various industries and businesses to aid in
repetitive processes. But there has been a burgeoning interest from
established tech giants and startups in using AI to make life a walk
in the park.

AI has been highly instrumental in optimizing the way we entertain

ourselves, interact with our mobile devices, to even driving vehicles
for us. We tend to encounter Machine Learning (ML) algorithms
and Natural Language Processing(NLP) in several everyday tasks
more than we know.

Two Categories for AI Usage

We can classify the way AI is utilized to improve the functionalities
of everyday life into two broad streams.

i)Software/Methodology:

Prominent examples of AI software used in everyday life include

voice assistants, image recognition for face unlock in mobile phones,
and ML-based financial fraud detection. AI software usually involves
just downloading software with AI capabilities from an online store
and requires no peripheral devices.

ii)Embodied:

The hardware side of AI includes its utilization in drones, self-driven

vehicles, assembly-line robots, and the Internet of Things (IoT). This
involves the design of specific devices that are based on AI
capabilities.
Question 1- Role of Ai in Human life/ daily life

10 Applications of AI Capabilities in Everyday

Life
AI and ML-powered software and devices are mimicking human
thought patterns to facilitate the digital transformation of society. AI
systems perceive their environment, deal with what they perceive,
solve problems and act to help with tasks to make everyday life
easier.

Here are 10 areas where AI plays an important role in helping

humans with their everyday requirements:

1)Voice Assistants

Digital assistants like Siri, Google Home, and Alexa use AI-backed
Voice User Interfaces (VUI) to process and decipher voice
commands. AI gives these applications the freedom to not solely
rely on voice commands but also leverage vast databases on cloud
storage platforms. These applications can then parse thousands of
lines of data per second to complete tasks and bring back tailored
search engine results.

There is a sweeping change in consumer awareness and an increase

in comfort levels with this technology. Voice assistant interfaces are
rapidly advancing, especially finding use in healthcare to identify
certain diseases through vocal biomarkers. Voice-based chatbots are
also being integrated into telehealth applications for triage and
screening.

2)Entertainment Streaming Apps

Streaming giants like Netflix, Spotify, and Hulu are continually

feeding data into machine learning algorithms to make the user
experience seamless.
Carefully looking at the user's interaction with various media, these
streaming apps recommend custom content. Using AI to parse
through the ever-increasing user data, these apps create catalogs of
music, movies, and TV series suited to each individual user's liking.

AI also plays an important role in providing uninterrupted streaming

by automating the allocation of servers closest to the user.
Bandwidth allocation also changes automatically based on the
popularity of a particular piece of media.

3)Personalized Marketing

Brands use AI-driven personalization solutions based on customer

data to drive more engagement. A report compiled by OneSpot
Research revealed that 88% of the surveyed consumers stated that
more personalized content makes them feel better about a brand.

Consumers are more likely to make a purchase when they are

directed to products through personalized marketing via automated
e-mailsand feedback forms. Recent innovations with AI claim to use
computer vision to predict an advertisement’s performance, helping
brands reach the right people and serving those who need those
products. Depending on the marketing stage, AI marketing apps
help both prospects and retargeted customers. Also, AI helps
businesses create a logo using a logo maker online tool that
resonates with the audience. It gives them insights into what
elements work best in their niche market or correctly portray their
products/services.

4)Smart Input Keyboards

The latest versions of mobile keyboard apps combine the provisions

of autocorrection and language detection to provide a user-friendly
experience.

With the help of AI, these apps can efficiently correct mistakes, help
switch between languages, and predict the next word in a non-
intrusive manner. Utilizing the principle of the "random forest"
machine learning algorithm, AI programmers are teaching these
apps to understand the context of the message being typed and
make accurate predictions.

Apps like Typewise and Swiftkey are now integrated with over 300
languages and dialects. Added facilities like real-time translation and
integrated search engines are also being introduced lately.
5)Navigation and Travel

The work of AI programmers behind navigation apps like Google

Maps and Waze never ends. Yottabytes of geographical data which
is updated every second can only be effectively cross-checked by ML
algorithms unleashed on satellite images.

Recently, researchers at MIT developed a navigation model that tags

road features in digital maps, all in real-time. These digital maps are
also created simultaneously based on satellite imagery incorporating
information about cycling lanes and parking spots.

Imaging algorithms built on Convolutional Neural Networks (CNN)

and Graph Neural Networks (GNN) have simplified regular updates in
routes. AI also helps ascertain routes on satellite images covered by
natural overgrowth with the help of predictive models.

6)Gamified Therapy

AI had found a place in gaming from the time classics such as Pac-
Man and Pong were around for intuitive universe-building. However,
until now innovations in gaming AI have focused on presenting more
interesting challenges to the gamer and not on gauging the gamer's
mindset.

Gamified applications are now being designed to gauge the gamer's

mental fortitude in the face of certain defeat. This is helping to study
methods to mitigate depression and anxiety in gamers and people in
general.

Using Virtual Reality (VR) headsets, some of these gaming

applications provide Cognitive Behavioral Therapy (CBT) for
enhanced engagement with the person. AI helps these games to
adapt to the user's behavior based on examined stimuli as the game
progresses.

7)Self-driving Vehicles

The technology of Autonomous Vehicle AI is witnessing large-scale

innovation driven by global corporate interest. AI is making
innovations beyond cruise-control and blind-spot detection to
include fully autonomous capabilities.

Deep Reinforcement Learning (DRL), a subset of machine learning,

is being applied to teach vehicles to operate independently. Path
planning in the face of static and dynamic obstacles is being made
possible through various predictive AI models.

Predicting accurately when neighboring vehicles will swerve and

other such unforeseen events are being taken into account.
Simultaneous Localization and Mapping (SLAM) is the technology
that makes this possible with real-time orientation to surroundings
via sensors.

8)Facial Recognition Technologies

The most popular application of this technology is in the Face ID

unlock feature in most of the flagship smartphone models today.
The biggest challenge faced by this technology is widespread
concern around the racial and gender bias of its use in forensics.

Generative Adversarial Neural Networks (GANN) are being applied to

reduce the margin of error in facial recognition software. These
neural networks are also being trained to pinpoint the unethical use
of Deepfake technology.

Various industries are also developing AI software that picks up

facial expressions to identify mood and intention. Emotion AI or
Affective Computing is an emerging area of interest to gauge
customer experience.

9)Security and Surveillance

It is nearly impossible for a human being to keep a constant eye on

too many monitors of a CCTV network at the same time. So,
naturally, we have felt the need to automate such surveillance tasks
and further enhance them by leveraging machine learning
methodologies.

AI frees up human surveillants to focus on the verification of crucial

incidents and act upon them. The constant monitoring and detection
part of surveillance is taken up by AI video monitoring software. AI
can pick up irregular behavior that might sometimes be missed by
human eyes.

An extension of AI-based facial recognition software is also being

employed in surveillance systems of high-risk public places like
government buildings. Currently, liberal governments are finding
ways to reduce the risk of privacy breaches from AI surveillance.

10)Internet of Things

The confluence of AI and the Internet of Things (IoT) opens up a

plethora of opportunities to develop smarter home appliances that
require minimal human interference to operate. While IoT deals with
devices interacting with the internet, the AI part helps these devices
to learn from data.

The five broad steps involved in IoT-enabling are - create,

communicate, aggregate, analyze and act. The efficacy of the
ultimate step, "act", is based on the depth of analysis, and AI adds
tons of value to it.

AI unlocks the potential of the data aggregated by IoT devices via

sensors. Learnings from this data over multiple iterations enable IoT
devices to respond to human stimuli and requirements better.

Question 2- Why Ethics in AI? Why Now?

The simple answer for this question is we need ethics which means logics
in Ai in certain to achieve optimal actions from an agent.

AI may involve robotics of varying complexity; it may involve the

manipulation of vast amounts of data; it may involve machine
learning. It may involve how we relate to others; it may involve how
as individuals we think, remember and reason. It may have
implications for the organisation of the labour market; it may involve
how we trade; who has access to information; and how. It has
implications for the economy, for politics, for culture. It has
implications for those who directly use AI, such as those who use a
robot butler; it has implications for those more remote, as for
example when an algorithm developed by machine learning helps
makes public policy decisions, or unemployment attributable to the
use of AI means a worker can no longer afford to pay their child’s
music teacher.

Artificial intelligence may be applied in many very different areas,

and in different ways within the same area. In medicine, AI may be
involved in computerised diagnosis of individual patients, or in
algorithms to analyse vast amounts of data from thousands or
millions of patients to understand the nature of disease and health.
It may be involved in patient consultations and even therapy
sessions with online or robotic responses. Robotic assistance with
surgery can involve AI, for complex and delicate operations. It may
be involved in remote monitoring of health, in mobile technology
that gives patients information about their own conditions. It may be
involved in nursing and care, with robotic assistants or companions.
Robots are being used to assist people with autism to develop social
skills. Robotic pets are being developed to provide companionship
and mental stimulation to patients with dementia. Robotic limbs are
being developed, as well as devices to enable patients with locked in
syndrome and other similar conditions to communicate.

Machinery increasingly involves AI. Autopilots include elements of

AI, and autonomous vehicles are imminently set for widespread use.
The ‘internet of things’ connects household gadgets and other items
using AI. Commerce involves AI, from automated trading agents in
the stock market, to algorithms which tailor online advertising or
sort the prices of airline tickets, both for buyers and for sellers. The
use of robotics in manufacturing is long established, and capabilities
are increasing. AI is moving into performing work that previously
required not just manual skills but intellectual skills, as in legal
research and accounting. Teaching also faces inroads from AI. AI
may even be used in creative endeavours, such in writing literature;
it’s already being used in the composition of music.

Question 3- Ethical considerations of Ai

1. How Do We Deal With Unemployment?

Let’s take, for example, the trucking industry, where millions of people
are employed in the United States alone. If Tesla’s Elon Musk delivers on
his promise of offering true self-driving cars (and by extension, delivery
trucks) and they become widely available within the next decade, then
what’s going to happen to those millions of people? But self-driving
trucks do seem like an ethical option when we consider their ability to
lower our accident rates.

2. How Can We Equitably Distribute The Wealth Created By

Machines?

Already, we are seeing startup founders take home the majority

of the economic surplus they generate. So how do we equitably distribute
the wealth created by machines?

3. Can Machines Influence Our Behavior And Interactions?

This challenge requires human raters to use text input to chat with an
unknown entity.
 4. How Do We Guard Against Possible Detrimental Mistakes?

we are to rely on AI to replace human labor, we need to ensure it

performs as planned and cannot be overpowered by humans with selfish
intentions.

5. Can We Eliminate AI Bias?

AI has the capability of speed and capacity processing that far exceeds
the capabilities of humans; however, due to human influence, it cannot
always be trusted to be neutral and fair.

6. How Do We Protect AI From Adversaries?

The more powerful the technology, the more it can be used for good as
well as various purposes.

Question 4-

Current initiatives in Ai and Ethics

Artificial intelligence (AI) has been rapidly advancing in recent years, and with it comes a growing
concern for the ethical implications of its use. As a result, there have been several initiatives aimed
at addressing these concerns and promoting responsible AI development. Here are some
examples:

1. The Partnership on AI: This is a collaboration between major tech companies such as Google,
Facebook, and Microsoft, as well as non-profit organizations and academic institutions. The
Partnership aims to promote responsible AI development and ensure that AI benefits society as a
whole.

2. The IEEE Global Initiative for Ethical Considerations in AI and Autonomous Systems: This
initiative is focused on developing standards and guidelines for the ethical use of AI and
autonomous systems. It includes stakeholders from various industries, including academia,
government, and industry.

3. The European Union's Ethics Guidelines for Trustworthy AI: These guidelines were developed by
the European Commission's High-Level Expert Group on AI and provide a framework for ensuring
that AI is developed and used in an ethical manner.

4. The Montreal Declaration for Responsible AI: This declaration was signed by over 1,000 AI
researchers and aims to promote the development of AI that is beneficial to society and respects
human rights.

5. The AI Now Institute: This is a research institute at New York University focused on studying the
social implications of AI. Its work includes analyzing the impact of AI on labor, civil rights, and
democracy.

These are just a few examples of the many initiatives underway to promote ethical AI
development. As AI continues to advance, it will be important to continue these efforts to ensure
that it is used in a way that benefits society as a whole.

Question -5
Ethical issues with our relationship with Artificial entities
As we develop more advanced artificial intelligence (AI) systems, there are several ethical issues
that arise in our relationship with these entities. Some of the key ethical concerns include:

1. Bias and Discrimination: AI systems can be biased based on the data they are trained on,
leading to discrimination against certain groups. For example, facial recognition software has been
shown to be less accurate for people with darker skin tones, which can lead to discriminatory
outcomes.

2. Privacy: AI systems can collect and analyze vast amounts of personal data, raising concerns
about privacy and surveillance. This is particularly concerning when AI systems are used by
governments or corporations with a lot of power.

3. Accountability: As AI systems become more autonomous, it becomes more difficult to hold them
accountable for their actions. This raises questions about who is responsible when an AI system
makes a mistake or causes harm.

4. Transparency: It can be difficult to understand how AI systems make decisions, which can lead
to a lack of transparency and accountability. This is particularly concerning when AI systems are
used in high-stakes decision-making contexts such as healthcare or criminal justice.

5. Job Displacement: As AI systems become more advanced, they have the potential to displace
human workers, leading to job loss and economic disruption.

6. Weaponization: AI systems can be used to develop autonomous weapons, which raises concerns
about the ethics of using machines to make life-and-death decisions.

7. Human-AI Interaction: As we interact more with AI systems, there are concerns about how this
will affect our relationships with each other and with the world around us. For example, some worry
that we will become overly reliant on AI systems and lose important skills and abilities.

These are just a few of the ethical issues that arise in our relationship with artificial entities. As AI
continues to advance, it will be important to address these issues in order to ensure that we use
these technologies in an ethical and responsible manner.

Unit-2
Framework and Models

Question 1- Elaborate Al Governance by Human-right centered design

Human-right centered design in AI governance refers to the development and

implementation of AI systems that prioritize human rights, dignity, and well-being. This approach
recognizes that AI systems have the potential to impact human lives in significant ways, and it is
important to ensure that these impacts are positive and aligned with our values.

One key aspect of human-right centered design in AI governance is ensuring that AI systems are
developed and deployed in a way that is transparent, accountable, and participatory. This means
involving a diverse range of stakeholders in the development process, including those who may be
impacted by the technology, and ensuring that there are mechanisms in place for ongoing
monitoring and evaluation.

Another important aspect of human-right centered design in AI governance is addressing issues of

bias and discrimination. This requires careful consideration of the data used to train AI systems, as
well as ongoing efforts to identify and address any biases that may arise. It also requires a
commitment to ensuring that AI systems are designed to promote equity and fairness, rather than
perpetuating existing inequalities.
Human-right centered design in AI governance also involves considering the impact of AI on human
dignity and autonomy. This means designing AI systems that respect individual rights and
freedoms, and that do not undermine our ability to make meaningful choices and decisions. It also
means considering the potential psychological and social impacts of AI on individuals and
communities, and taking steps to mitigate any negative effects.

Finally, human-right centered design in AI governance requires a commitment to ongoing reflection

and learning. As AI technologies continue to evolve, it is important to regularly assess their impact
on human rights and well-being, and to make adjustments as needed to ensure that they align with
our values and priorities.

Overall, human-right centered design in AI governance is essential for ensuring that AI systems are
developed and used in a way that promotes human flourishing and well-being. By prioritizing
human rights and dignity in the development of AI technologies, we can harness their potential for
positive impact while mitigating any negative consequences.

Question- 2 NORMATIVE MODELS

Normative models in ethics of Artificial Intelligence (AI) refer to the frameworks and principles that
guide the development and use of AI systems in a way that aligns with ethical values and norms.
There are several normative models in ethics of AI, including consequentialism, deontological
ethics, virtue ethics, and care ethics.

Consequentialism is a normative model that evaluates the morality of an action based on its
consequences. In the context of AI, consequentialism would prioritize the development and use of
AI systems that produce the greatest good for the greatest number of people. This model may be
used to justify the use of AI for tasks such as medical diagnosis, disaster response, and
environmental monitoring.

Deontological ethics is a normative model that emphasizes the importance of following moral rules
and duties. In the context of AI, deontological ethics would prioritize the development and use of AI
systems that respect human rights and dignity, regardless of their potential consequences. This
model may be used to justify the prohibition of certain uses of AI, such as autonomous weapons or
mass surveillance.

Virtue ethics is a normative model that focuses on developing virtuous character traits, such as
compassion, honesty, and courage. In the context of AI, virtue ethics would prioritize the
development and use of AI systems that promote these character traits in individuals and
communities. This model may be used to justify the use of AI for tasks such as education, mental
health support, and social services.

Care ethics is a normative model that emphasizes the importance of relationships and empathy in
moral decision-making. In the context of AI, care ethics would prioritize the development and use
of AI systems that foster human connection and social cohesion. This model may be used to justify
the use of AI for tasks such as elder care, childcare, and community building.

Overall, normative models in ethics of AI provide a framework for evaluating the moral implications
of AI development and use. By considering the ethical principles and values that underlie these
models, we can ensure that AI systems are developed and used in a way that aligns with our
ethical commitments and promotes human flourishing and well-being.

Question -3 Role of professional norms

Professional norms play a crucial role in the ethics of artificial intelligence (AI) by providing
guidance and standards for the development and use of AI systems. These norms are developed
and enforced by professional organizations, such as the Association for Computing Machinery
(ACM) and the Institute of Electrical and Electronics Engineers (IEEE), which have established codes
of ethics for their members.

Professional norms in AI ethics typically include principles such as transparency, accountability,

fairness, and privacy. These principles reflect the ethical values and norms that are widely
recognized in society and are essential for ensuring that AI systems are developed and used in a
way that aligns with these values.

For example, the principle of transparency requires that AI systems be designed and operated in a
way that is understandable and explainable to users and stakeholders. This principle is important
for ensuring that AI systems are not used in a way that is discriminatory or biased, as it allows for
the identification and correction of such issues.
Similarly, the principle of accountability requires that individuals and organizations responsible for
the development and use of AI systems be held responsible for their actions. This principle is
important for ensuring that AI systems are used in a way that is ethical and legal, as it provides a
mechanism for holding individuals and organizations accountable for any harm caused by their
actions.

Overall, professional norms play a critical role in the ethics of AI by providing guidance and
standards for the development and use of AI systems. By adhering to these norms, professionals in
the field of AI can ensure that their work aligns with ethical values and norms, promoting human
flourishing and well-being.

Question-4 Moral machines

Teaching a machine to be moral in AI involves several approaches, including:

1. Developing ethical frameworks: It is essential to develop ethical frameworks that guide the
behavior of AI systems. These frameworks should be based on ethical principles and values that
are widely recognized in society, such as fairness, transparency, and accountability.

2. Incorporating moral reasoning: AI systems can be designed to incorporate moral reasoning,

which involves making decisions based on ethical principles and values. This can be achieved
through machine learning algorithms that are trained on ethical scenarios and principles.

3. Building in feedback mechanisms: AI systems can be designed to incorporate feedback

mechanisms that allow them to learn from their mistakes and adjust their behavior accordingly.
This can help to ensure that the system's actions align with ethical principles and values.

4. Collaborating with ethicists: Collaboration with ethicists can help to ensure that AI systems are
developed and used in a way that aligns with ethical principles and values. Ethicists can provide
guidance on ethical issues related to AI development and use.
5. Conducting regular audits: Regular audits of AI systems can help to identify any ethical issues or
biases in the system's behavior. This can help to ensure that the system's actions align with ethical
principles and values.

Overall, teaching a machine to be moral in AI involves a combination of approaches, including

developing ethical frameworks, incorporating moral reasoning, building in feedback mechanisms,
collaborating with ethicists, and conducting regular audits.

Example
One example of teaching AI machines to be moral is in the development of autonomous vehicles.
In situations where a collision is unavoidable, the AI system must make a decision about who to
prioritize for safety - the passengers in the vehicle or pedestrians on the road. By incorporating
moral reasoning and ethical principles into the decision-making process, the AI system can make a
decision that aligns with societal values and reduces harm to all parties involved. This involves
developing ethical frameworks, building in feedback mechanisms, and collaborating with ethicists
to ensure that the system's actions align with ethical principles and values.

Unit – 3
Concepts and issues

Question-1 Accountability in computer systems

Accountability in computer systems refers to the responsibility of individuals or organizations for

the actions and decisions made by the systems they create and operate. As AI and other advanced
technologies become more prevalent in our daily lives, it is increasingly important to ensure that
these systems are accountable for their actions and decisions.

There are several ways in which accountability can be built into computer systems. One approach
is to ensure that the system is transparent, meaning that its decision-making process is visible and
understandable to users and stakeholders. This can be achieved through techniques such as
explainable AI, which allows users to understand how a system arrived at a particular decision.

Another approach is to incorporate feedback mechanisms into the system, allowing users to
provide input on the system's performance and decision-making. This can help to identify and
address potential biases or errors in the system.
In addition to technical solutions, accountability also requires a legal and regulatory framework.
This includes laws and regulations that govern the use of AI and other advanced technologies, as
well as mechanisms for holding individuals and organizations accountable for their actions.

Overall, accountability is critical for ensuring that computer systems are used responsibly and
ethically, and that they align with societal values and goals. By building accountability into these
systems, we can help to mitigate the risks and maximize the benefits of these technologies for
individuals and society as a whole.
Question 2- Elaborate the following concepts of Artificial intelligence ethics

1.Transparency. 2.Responsibility and Al. 3.Race and Gender.

4.Al as a moral right-holder.

Answer-

1. Transparency:

a) Importance: Transparency is crucial in AI systems as it allows users to understand how the

system arrived at a particular decision. This understanding is essential for building trust in the
system and ensuring that the decisions made by the system are fair and unbiased.

b) Challenges: Achieving transparency in AI systems can be challenging as many AI systems use

complex algorithms that are difficult to understand. Additionally, some organizations may be
reluctant to share information about their AI systems, fearing that it could reveal trade secrets or
other sensitive information.

c) Solutions: Techniques such as explainable AI can help achieve transparency in AI systems.

Explainable AI allows users to understand the reasoning behind a system's decision-making
process. Additionally, organizations can be encouraged to share information about their AI systems
through regulations and incentives.

2. Responsibility and AI:

a) Importance: As AI systems become more prevalent in our daily lives, it is important to ensure
that individuals and organizations are held responsible for their actions and decisions. This includes
ensuring that AI systems are designed and used in an ethical and responsible manner, and that
individuals and organizations are accountable for any negative impacts of these systems.

b) Challenges: Holding individuals and organizations responsible for the actions of AI systems can
be challenging as it can be difficult to determine who is ultimately responsible for the system's
actions. Additionally, there may be a lack of regulations or guidelines governing the use of AI
systems.

c) Solutions: Regulations and guidelines can help ensure that AI systems are designed and used in
an ethical and responsible manner. Additionally, organizations can be encouraged to develop
internal policies and procedures for the use of AI systems, including measures for accountability
and transparency.

3. Race and Gender:

a) Importance: AI systems have the potential to perpetuate or even amplify existing biases and
discrimination in society. For example, facial recognition technology has been shown to have
higher error rates for people of color and women. It is important to consider issues of race and
gender in the design and use of AI systems, and to ensure that these systems are not perpetuating
or amplifying existing biases and discrimination.

b) Challenges: Addressing issues of race and gender in AI systems can be challenging as biases
can be difficult to identify and address. Additionally, there may be a lack of diversity in the
development teams creating AI systems, which can lead to blind spots in the system's design.

c) Solutions: Increasing diversity in development teams can help address blind spots in the design
of AI systems. Additionally, testing and auditing can be used to identify and address biases in AI
systems. Finally, regulations can be put in place to ensure that AI systems are designed and used
in a way that does not perpetuate or amplify existing biases and discrimination.
4. AI as a moral right-holder:

a) Importance: The concept of AI as a moral right-holder suggests that AI systems may have moral
rights similar to those of humans. This concept is important as it raises questions about how we
should treat AI systems and whether they should be protected from harm and exploitation.

b) Challenges: The concept of AI as a moral right-holder is highly debated, and there is no

consensus on whether AI systems can truly be considered moral right-holders. Additionally, if AI
systems were given moral rights, it could have significant legal and ethical implications.

c) Solutions: Further research and debate are needed to determine whether AI systems can truly
be considered moral right-holders. Additionally, regulations and guidelines can be put in place to
ensure that AI systems are treated with respect and dignity, regardless of whether they are
considered moral right-holders.