Page 284 - Touhpad Ai

● Explainability: Explainability in AI is crucial for ensuring that the decisions made by AI systems are
understandable to humans. It pertains to the transparency and clarity of AI systems, enabling users
to understand the decision-making process and forecasts of algorithms.
● Robustness: Robustness in AI ethics refers to the capacity of AI systems to perform reliably and
accurately across various conditions, while also minimising unintended consequences and harmful
impacts. It is a fundamental aspect of ethical AI because unreliable or biased systems can lead to
significant societal harm.

● Transparency: Transparency in AI means being open and clear about how AI systems are created,
how they work, and what impacts they might have. It involves providing straightforward information
about the data, algorithms, and decision-making processes used in AI applications. This openness
encourages accountability, allows for scrutiny, and helps people make informed choices about the
ethical and social implications of AI technologies.
● Privacy: Privacy involves individuals having control over their personal information and avoiding
unwarranted interference in their lives. It encompasses the right to keep aspects of one's life
private, such as private messages, activities, and data. Privacy is crucial as it safeguards individual
autonomy, dignity, and freedom from unnecessary intrusion.

2. Discuss the concept of the ethical dilemma with an example.
Ans. An ethical dilemma is a situation in which a person faces a choice between conflicting moral principles
or values. It often involves tough decisions where there are competing interests or where doing what is
considered right may result in undesirable outcomes. Ethical dilemmas can arise in various contexts,
such as in personal relationships, professional settings, or societal issues. Resolving ethical dilemmas
requires thoughtful consideration of the consequences of different actions and balancing conflicting
ethical concerns.
Let us understand the concept of ethical dilemma with the help of an example.

● Scenario: You work for a pharmaceutical company developing a new drug to treat a rare disease.
During clinical trials, it becomes evident that the drug is effective in treating the disease, but it also
has significant side effects in a small percentage of patients. The company is under pressure to
release the drug quickly due to the urgent need for treatment, but there are concerns about the
potential harm caused by the side effects.
● Ethical Dilemma: On one hand, releasing the drug could provide relief to patients suffering from the
rare disease, potentially saving lives, and improving quality of life. On the other hand, there’s a risk
of causing harm to patients due to the side effects, which could lead to serious health complications
or even fatalities.
3. Explain the different sources of bias in AI systems and how they can lead to unfair outcomes.
Ans. Bias in AI systems can stem from several sources, including training data bias, algorithmic bias, and
cognitive bias.
● Training data bias: This occurs when the data used to train AI systems is unrepresentative,
incomplete, or skewed. For instance, if a medical AI system is trained primarily on data from male
patients, it may not perform well for female patients, leading to misdiagnoses. Similarly, an AI used
for loan approvals might be biased if it primarily includes applicants from affluent neighbourhoods,
thereby ignoring applicants from poorer areas.

● Algorithmic bias: This type of bias arises during the design and implementation of algorithms. If
an AI hiring algorithm is trained on historical data that reflects biased hiring decisions, such as
favouring one demographic group over another, the algorithm may perpetuate these biases in new
hiring recommendations.

282 Touchpad Artificial Intelligence - XI