AI Ethics, Bias and Safety

AI is a powerful tool, so it is important to think about how it can harm people and how to use it fairly.

Bias: When AI is Unfair

AI learns from data. If the data is unfair, the AI learns the unfairness too. This is called bias.

Example: a company trains a hiring model on its past hires. If most past hires were men, the model may learn to prefer men, even if women are equally qualified.

Here is a tiny example that shows how skewed data gives skewed results:

past_hires = ["male"] * 18 + ["female"] * 2
share_female = past_hires.count("female") / len(past_hires)
print("Female share in training data:", share_female)
if share_female < 0.3:
    print("Warning: data is unbalanced. The model may be biased.")

Female share in training data: 0.1
Warning: data is unbalanced. The model may be biased.

Checking the balance of your data is one of the first steps to building fairer AI.

Main Ethical Concerns

Privacy

Many chatbots may store your conversations, and some use them to improve their models unless you turn this off. So:

• Never paste passwords, CNIC numbers, bank details or private documents into a public AI tool.
• Do not share other people's personal information without their permission.
• Read the privacy settings of the tool you use.

Deepfakes and Scams

AI can copy a person's voice from a short recording. Scammers have used fake voice calls pretending to be a relative asking for money urgently. Protect yourself:

Safety checklist for a suspicious call or video:
1. Is it asking for money or codes urgently?
2. Can you call the person back on their known number?
3. Ask a question only the real person would know.
4. Never share OTP codes, even with "the bank".

AI Safety

AI safety is the work of making sure AI systems do what people intend and do not cause harm. AI companies test their models, add rules to refuse dangerous requests and study how models behave. Governments in many countries, including Pakistan, are writing AI policies and laws.

As AI agents begin to take actions (send emails, make purchases, change files), it becomes more important that a human approves important steps.

Using AI Responsibly: Simple Rules

1. Check facts before sharing anything an AI tells you.
2. Be honest: follow your school or employer's rules about AI, and say when work is AI-assisted if required.
3. Protect privacy: yours and other people's.
4. Do not create harmful content: no fake images of real people, no harassment.
5. Keep a human in charge of decisions about health, money, jobs or the law.

Exercise