Why AI Needs Ethics

A calculator cannot be unfair. It adds, and that is all. An AI is different: it decides who gets a loan, which patient is seen first, whose application is read. A thing that decides about people can be unfair, and usually without meaning to be.

1Ethics, then ethical frameworks

Ethics

Ethics are a set of values or morals which help us separate right from wrong.

Ethical framework

Frameworks are step-by-step guidance for solving problems. So an ethical framework is a framework which helps us ensure that the choices we make do not cause unintended harm.

Put the two definitions side by side and the phrase writes itself: ethics tells you right from wrong, a framework gives you steps, and an ethical framework is the steps that keep your choices on the right side.

2Why AI in particular

AI is being used as a decision-making and decision-influencing tool. Because of that, we have to make sure it makes morally acceptable recommendations. Ethical frameworks are how: if we use them while building a solution, we can avoid unintended outcomes even before they take place.

Bias, and where it comes from

A model learns only from its data. A famous hiring algorithm was trained on a company’s past hires, most of whom were men — so it learned to reject women. Nobody wrote that rule. The data contained it, and the model copied it faithfully. This is bias, and it is the commonest way an AI causes harm without anyone intending it.

A mirror, not a judge

A model does not have opinions. It reflects what it was shown. If what it was shown was unfair, the reflection is unfair — and because it comes out of a computer, people believe it more, not less.

3Before, not after

The point of a framework is that it is used while you are building. Checking a finished system for harm is better than nothing, but by then people have already been affected. Below is what using one looks like: a recommendation has to pass every check before it is allowed to reach a patient.

Four gates on the way to the patient

Each gate is one principle of bioethics. A decision that fails a principle does not get to trade it off against the others — it stops there.

The recommendation cannot get through. It fails Do Not Harm — patients who were sicker were sent less care — and it fails Give Justice, because one region carried all of that harm.

1. Respect for Autonomy

People must be able to decide for themselves, which means they must know how the AI reached its decision.

2. Do Not Harm

Harm to anyone — human or non-human — must be avoided. Where no harmless choice exists, take the path of least harm.

3. Ensure Maximum Benefit

Avoiding harm is not enough. The action should give the greatest possible benefit to everyone involved.

4. Give Justice

Benefits and harms must be shared fairly. No group may be left worse off than another.

4From the board papers

Ethical frameworks are primarily designed to:
(A) Increase the efficiency of AI algorithms (B) Ensure that choices made do not cause unintended harm (C) Reduce the cost of AI development (D) Speed up the AI project cycleCBSE board exam 2026

1 mark

▶Show the answer

(B). This is almost word for word the definition in your book. An ethical framework is not about speed, cost or efficiency — it is about harm.

5Check yourself

Quick Check

What are ethics?

Quick Check

A hiring model trained on a company's past hires rejects almost all women. What went wrong?

← Previous

🪜 What is a Framework?

Types of Ethical Frameworks 🗂️