Page 73 - CT_AI_Class-7
P. 73
The Reserve Bank of India has also set rules for digital lending fact bits
platforms that use AI to make decisions about loans and credit. From 20 February 2026,
th
These rules make sure the platforms are fair and transparent to under India’s amended
customers. For example, if you apply for a loan online, the system Information Technology Rules,
uses AI to decide whether to approve it. Ethical AI ensures that 2026, social media platforms
the AI system is fair and transparent, so everyone has an equal must remove harmful content
chance to get a loan. within 2 hours of a complaint.
BIAS IN AI
Bias in AI means unfair decisions that favour some people over others. AI doesn’t have its own
biases, but it can learn them from the data it’s trained on. If the data is unfair, the AI can make
unfair decisions. Here are a few examples of bias in AI:
Inaccurate descriptions: Imagine an AI that recommends books, but it’s trained only on books
for boys. It might only suggest books for boys and ignore books for girls. This isn’t fair, as girls
may enjoy other topics like mystery or adventure. Ethical AI should recommend books based
on everyone’s interests.
Inaccurate assumptions: If an AI system mostly sees data about people who like football, it
might assume everyone likes football. This could lead to biased suggestions, missing out on
people who enjoy other sports like cricket or basketball.
Wrong data: If an AI system is trained with data from one place, it might make mistakes when
used in other places. For example, if the system is trained only with data from one city, it might
not understand the needs of students in different cities.
For instance, if an AI system trained only on data from one gender or one geographical location
is used universally, it might make unfair suggestions or decisions. This is why it’s important to
ensure AI systems are trained on balanced data and are carefully designed to avoid perpetuating
bias. Bias can happen for three main reasons:
The data used: If the data isn’t balanced, the AI can learn biased patterns.
How the system is designed: Sometimes the system itself can unintentionally favour one group.
Human decisions: The people who create the AI system can pass their own biases onto it.
Ethical AI makes sure the system is fair and treats everyone equally.
UNDERSTANDING THE MEANING OF BIAS
Bias means treating people unfairly or making decisions that aren’t based on all the facts. It
happens when someone prefers one person or thing over another for no good reason. In simple
terms, bias is when things aren’t done fairly. For example, imagine your school’s basketball coach
is picking players for the team. He only selects tall students without considering their skills. But
what if a shorter student is really good at basketball? That’s unfair. The coach’s decision is biased
because it ignores the abilities of shorter students.
Ethics and AI Bias Awareness 71
