Page 331 - Artificial Intellegence_v2.0_Class_11
P. 331
Establish a Mechanism of Global Governance
Ethical AI governance bodies on global and regional levels need to be established. These bodies must include all
stakeholders—AI designers, manufacturers, owners, developers, researchers, employers, lawyers, CSOs, and trade unions.
Prohibition of Assigning Responsibilities to Robots
The design and operation of robots should comply with existing laws and fundamental rights and freedoms, including
privacy, as much as possible.
Prohibit AI Arms Race
Lethal autonomous weapons, as well as cyber warfare, should be prohibited.
Reboot
Give any two examples where AI is helping people.
____________________________________________________________________________________________________________________
____________________________________________________________________________________________________________________
What is Bias?
Bias is any prejudice against individuals or groups, especially in ways that are considered unfair. "Bias in AI" has long been
a key area of research and attention in the machine learning community. It is a term used to describe situations where
ML-based data analysis systems are biased against certain groups of people. These prejudices usually reflect the prevailing
social prejudices about race, gender, biological sex, age, and culture. AI systems learn to make decisions based on training
data, which may include biased human decisions or reflect historical or social inequities.
Types of Bias and How It Influences AI Based Decisions
The most popular classification of bias in artificial intelligence divides it into three categories: algorithmic, data, and human.
This classification uses the source of prejudice as the primary criterion. However, as human bias is the root of and exceeds
the other two, researchers and practitioners of AI urge to watch out for the latter. The most typical forms of AI bias that
seep into the algorithms are listed below:
Reporting
Types
Implicit Selection
of Bias
Group
Attribution
AI Values (Bias Awareness) 329

