SUBJECT: ARTIFICIAL INTELLIGENCE

CLASS : IX
UNIT –1
Introduction to Artificial Intelligence
Handout 2

CHATBOTS
A chatbot is an artificial intelligence (AI) software that can simulate or imitate a conversation
(or a chat) with a user in natural language through messaging applications, websites, mobile
apps or through the telephone.
In 1966, ELIZA was the first natural
language processing (NLP) based
conversation program described by
Joseph Weizenbaum.It featured a
conversation between a human user and
a computer program representing a
mock psychotherapist.
As you interact more and more with chatbots, you would realise that some of them are scripted
or in other words are the traditional chatbots while others are AI powered and have more
knowledge. There are two types of chatbots:
1. Script bots - These are the chatbots which work around a script which are
programmed in them and they can handle the conversations only. They work on
pre-written keywords that they understand. Each of the commands that they are going
to follow must be coded into them by the developer. So, if a user asks them something
outside of their knowledge base, they respond with “Sorry, I did not understand”.
Example : The story speaker bot, the bots which are deployed in the customer care
section of various companies(like amazon, zomato, flipkart etc.)
2. Smart bots - These are the chatbots which not only can handle the conversation but
can also manage to do other tasks which make them smarter.Based on AI, these bots
don’t have pre-programmed answers. They learn with time, catching keywords and
putting them in context, and help users arrive at the most relevant answers to their
queries.
Example : All the assistants like Google Assistant , Alexa, Cortana, Siri etc.
Reasons why customers prefer chatbots are as follows:
○ Productivity: Chatbots provide fast and efficient access to information thereby,
improving productivity.
○ Entertainment: Since chatbots are AI-powered, they can hold amusing conversations
with people while they are waiting on a platform. This also helps keep users engaged
with the site.
○ Social and relational factors: Fueling conversations, chatbots enhance social
experiences of visitors. Not only chatbots help users to speak up without being judged,
they also help them enhance their conversation skills.
○ Curiosity: People are always curious about trying something new. And the novelty of
 chatbots sparks amazing curiosity.

UNDERSTANDING AI ETHICS
In the simplest possible words, we can define ‘ethics’ as a system of moral principles that
govern an individual's behaviour and actions. Ethics are concerned with what is good for
individuals and societies. Similarly, ethical concerns are the issues, situations or the
concerns that cause individuals societies and organisations to evaluate different choices in
terms of what is right(ethical) and what is wrong(unethical).
At the core of the AI challenges lies the problem of existing human biases which will enter
AI systems through both the developers and the data. Bias can be of two type -
1. Model Bias - The model is adequately represent the dataset accurately enough
2. Prejudice Bias - The creator’s stand on certain issues is embedded intentionally or
unintentionally in the algorithm.
BIAS IN REAL WORLD DATA
The problem here is that an AI system learns from the real world data fed into it. This means
that a system can reinforce the biases found in AI systems. For example, a computer system
trained on the data for the last 200 years might find that the more females were involved in
certain jobs or that more percentage of successful businesses were established by men and
conclude that specific genders are better equipped for handling certain jobs(gender bias).
Understanding or even detecting such biases is not easy because many AI systems act as
black boxes i.e. the reason behind their decision making is not easy or in some cases, even
possible to understand. Many times, programmers of AI systems themselves cannot explain
the logic behind decisions taken by the AI systems. Let's look at two famous cases of AI
Bias.

The only way through which trust in AI systems can be built up is explainability. Let us
understand what it exactly means. DARPA, US, initiated a project called explainable AI
(XAI) program with the following aims-
● Producing more explainable AI models while maintaining a high level of learning
performance
● Augmenting human intelligence and decision-making, enabling humans to understand
 and trust appropriately.
● The machine learning algorithms need to have the inbuilt capability to explain their
logic, define their strengths/weaknesses and specify a clear understanding of the future
behaviour.
Problem of Inclusion
AI systems trained on biased real world data create the problem of inclusion, i.e. the problem
that some people are left out of their decision making system. Consider the example of an AI
system used by Amazon for recruitment. This created a situation in which many eligible
females were left out of consideration. This is known as the problem of inclusion.
The root cores of the bias in AI systems come from the problem between facts and
interpretation of facts. AI systems can scan data and draw learnings from them but these
systems are not equipped to understand the reason behind particular conclusion or learning
Microsoft developed and launched an experimental Twitter chatbot known as ‘Tay’. It was
designed to experiment with conversational understanding. While launching Tay, Microsoft
claimed that, the more you chat with it, the smarter it will get, learning to engage people
through casual and playful conversations. In less than 24 hours of its launch, Tay learnt from
thousands of tweets and developed a racial misogynistic personality.
There was no programming error or hacking for the consequences. The problem was that Tay
did exactly what it was designed to do. It started mimicking the users who were talking to it.
It learned from what was being fed to it without interpreting the inputs.
Concerns related to implications of the AI technology itself.
The second set of ethical problems is related to the implication of the AI systems. This AI
ethical concerns can be divided into three parts
1. Privacy concerns of AI system
2. Ethical concerns related to adoption of AI systems
3. Security in AI
Privacy concerns of AI system
One of the most serious implications of data collection is related to privacy concerns. Take a
simple example of your Android smartphone. The following is a list of some information that
Google possesses about you (if you have given it permission)-
1. Contact list
2. Location
3. Email Id
4. Hangout chat records
5. Photos - The images of you and your friends along with their names, their phone numbers
and all other details like birthdays, anniversaries etc.
Most of the time we accept lengthy user agreements without even realising the implications
of this agreement on our privacy rights. Data related to us is stored somewhere in some large
databases. This data can be used for a number of purposes. The information is collected in
lieu of providing us with the helpful services, but the potential risk here is very high. For
example, take the facial recognition system. The AI system can use this technology along
with the information provided by the apps like contact and photos to create a wide
surveillance network that can monitor you 24X7.
Ethical concerns related to adoption of AI systems
The major concerns related to the adoption of AI systems is
1. Job loss
2. Increasing inequalities
3. Negative adoptions
4. Black box problem
Security in AI
Everything logical development can be used for the benefit of mankind as well as for its
destruction. We can use dynamites to make mining easy and we can also use them as
weapons. Nuclear power can be used to create energy and it can also be used to destroy cities
and communities. Advancement in medicines can be used to heal and improve the quality of
life of human beings as well as to create biological weapons. It can be used for the good of
mankind or its destruction.
AI systems need to be secure and reliable if their usage has to be increased. We need to
understand that every new technology goes through a transition and it takes time before it
becomes reliable.
The key solutions around reliability could be:
● Testing guidelines of aI systems before being used by humans
● Liability of AI systems for any consequences
● Defining AI consequences and comparing with what a human would do in that scenario to
define gaps. Don't expect miracles from AI!
Overall, society will have to come up with ethical guidelines on data and AI systems as it has
done for many of the other technologies. The violation of these guidelines, rules or
regulations will have to be punished appropriately.
The United Kingdom is taking the lead in this direction by investing 9 million pounds to
establish the centre for Data Ethics and Innovation. This centre will ensure ethical, safe and
innovative uses of data. This may also include possibilities of establishing data trusts to
facilitate easy and secure sharing of data. A consortium of Ethics Councils may be formed to
define the standard practice. It would be expected that all Centres of Excellence adhere to
standard good ethical practices while developing AI technology and products.
SUMMARY OF WHAT ALL IS COVERED UNDER ETHICS-
Basic of AI Ethics -
● Bias, Prejudice, Fairness
● Who is accountable
● Transparency, Interpretability, Explainability
Actions of AI
● Safety
● Human AI Interaction
● Cybersecurity and Intentional Misuse
● Data Privacy and Control
AI Impact -
● Job Losses and Unemployment
● Civil Rights - Robot Rights
● Human-Human Interaction Change
● Economy