Page 43 - 2617_JSSPS_C-7

What if AI Makes Mistakes: Yes, it's true, AI can make mistakes. It's recorded that in 2016, when the
Uber company conducted a test on self-driving cars in San Francisco, Uber’s autonomous vehicle ran
six red lights. The situation got out of hand, but a licenced driver was made to sit behind the wheel in
case of emergency so he took over control of the situation immediately. In another situation in the
same year, Microsoft’s chatbot, Tay was released on Twitter. It is based on machine learning, natural

language processing and social networks. It learnt its language from the people on Twitter over time
that enabled it to have a meaningful conversation based on a topic. But sadly, it was taken offline
within 16 hours of its launch as it started tweeting randomly some abusive and offensive content.

ETHICAL PILLARS OF AI

To make AI systems more accurate, fair and trustworthy, it is essential to recognise and address the
factors that cause these biases. One effective way to achieve this is by integrating ethics into AI design
and development. When ethical principles are included from the very beginning, AI can be built to
respect fairness, transparency and equality, ensuring it benefits all people—not just a few.

Privacy Transparency Explainability Robustness Fairness

AI systems The way AI should AI should be AI must treat
must protect AI makes provide clear safe, reliable all people
personal decisions reasons for and able to equally and
data and should be its actions or handle errors avoid bias or
avoid misuse open and decisions. or unexpected discrimination.
of user easy to situations.
information. understand.

5 Pillars of AI Ethics

Because doing so ensures that AI systems are:

Human-centric design: Ethical AI places human well-being, dignity and values at the centre of
technological development. It ensures that AI systems are designed to serve people, support their
rights and meet their real-world needs. This involves creating user-friendly, inclusive and accessible
solutions that enhance both individual experiences and broader societal benefits.

Fair and unbiased: One of the key goals of ethical AI is to reduce and prevent bias in algorithms
and decision-making processes. This means actively working to eliminate discrimination based on
race, gender, age or socioeconomic background. By promoting fairness, equity and inclusivity, AI can
support just and balanced outcomes for all users.

Fields & Future of AI 41