Page 79 - CT_AI_Class-8

 Test systems carefully before deployment: Before an AI system is used in real-world situations,
it must be tested thoroughly. Testing helps identify unfair patterns, errors or weaknesses in the
system. Developers can then fix these issues before the system is widely used.

For example, if an AI system used for loan approval is tested with data from different income
groups and regions, developers can check whether the system treats all applicants fairly.
 Allow human review of important decisions: Some decisions made by AI systems can have

a significant impact on people’s lives. In such cases, human review is important. Humans can
check the system’s decisions and correct them if necessary.
For example, if an AI system rejects a candidate during job screening, a human recruiter can
review the application again to ensure that no deserving candidate is unfairly rejected.

 Monitor performance regularly: AI systems should be monitored even after deployment.
Over time, new types of data and situations may arise. Continuous monitoring helps detect
new biases or errors and ensures that the system continues to perform fairly.

For example, a recommendation system used in online learning platforms should be monitored
regularly to ensure that it recommends content fairly to all students and does not favour only
certain groups.

 Update models with new data: AI systems improve when they learn from new and updated
data. Regular updates help the system adapt to changing conditions and reduce bias.
For example, a navigation system should be updated with new road information, traffic
patterns and locations. This helps the system provide accurate and fair suggestions to users
from different areas.

A titanium-alloy smart ring designed for continuous health
monitoring, including heart rate, blood oxygen levels and
sleep quality. It uses AI to analyse biometric data, provide
personalised wellness insights and give early alerts for stress or
ai
fatigue.

lens

Importance of Fairness in AI

Fairness is an essential principle in Artificial Intelligence because AI systems increasingly
influence decisions that affect people's lives. These decisions may relate to education, healthcare,
employment, finance and many other areas. If AI systems are not fair, they may create unequal

opportunities and negatively impact certain individuals or groups.
Some importance of fairness in AI system is as follows:

 Ensures equal treatment for all individuals, regardless of gender, age, region or background
 Reduces discrimination and prevents biased decision-making

 Builds trust and confidence in AI systems among users

Ethics and Responsible AI 77