The guide names bias and fairness, accountability and transparency, and the erosion of human judgement. They are different problems and they need different answers — which is why an essay that treats them as one thing stalls in the middle band.
One decision reaching a person, and the four separate things accountability needs.
Interactive diagram
Explore the labelled diagram, charts and maps for this topic in full study mode.
Free preview
This is the free notes preview
You're reading the free notes. Aimnova Pro unlocks the full study experience — and you can try it with your first topic free to keep:
- FlashcardsLock in vocabulary and key terms with spaced repetition.
- Practice questionsAnswer exam-style questions and get instant AI marking.
- Mock exams & past-paper vaultSit full mocks and see exactly how examiners award marks.
- Personalised study planA daily plan built around your exam date and weak areas.
This is the point students most often miss. A system can be built by careful people, with no rule about any protected group, and still produce unfair outcomes.
Where the bias comes from instead
- The training data. It records what happened before, including decisions nobody would defend now.
- Who is in the data. A group that was rarely recorded is rarely predicted well.
- The stand-in measure. Using postcode for reliability, or a career gap for commitment, imports whatever those things track.
- The target itself. “Who got promoted before” is not the same question as “who would do the job well”.
The sentence to learn: Nobody wrote the rule. The data taught it. That is what distinguishes algorithmic bias from ordinary discrimination, and it is why removing the obvious fields does not fix it.
Know your predicted grade
Take timed mock exams and get detailed feedback on every answer. See exactly where you're losing marks.
| Dilemma | The question it asks | Why supplying only this is not enough |
|---|---|---|
| Transparency | Can the reasoning be seen? | Knowing why changes nothing if nobody must act on it |
| Accountability | Is there someone who must answer? | A duty to answer is empty if the reasoning is secret |
| Black box | Can the reasoning be seen by ANYONE? | Some models have no explanation to give, even to their makers |
| Erosion of judgement | Did a person really decide? | A human sign-off on a score nobody can check is not a review |
Real-world examples you can name
COMPAS recidivism scoring — ProPublica investigation May 2016
A commercial risk-scoring tool used in some US courts to estimate how likely a defendant was to reoffend. Journalists found that among people who did not go on to reoffend, Black defendants were roughly twice as likely as white defendants to have been labelled high risk. The company disputed the measure of fairness used, and the argument that followed showed that competing definitions of a fair algorithm cannot all be satisfied at once.
Who it affected: Defendants whose bail and sentencing decisions were informed by the score.
Amazon's scrapped CV-screening tool — reported October 2018
An experimental hiring tool was trained on a decade of CVs submitted to the company, most of them from men. It learned to downgrade CVs containing the word 'women's' and graduates of two women's colleges. The project was abandoned.
Who it affected: Women applying for technical roles, and every organization that assumed historical data was neutral.
The EU Artificial Intelligence Act — agreed December 2023; entered into force August 2024, phased to 2027
The first broad law to regulate AI by risk. Some uses are banned outright, such as untargeted scraping of facial images and social scoring by public authorities; high-risk uses in areas like hiring and education carry duties for data quality, documentation and human oversight.
Who it affected: Anyone deploying AI systems in the EU, and users subject to those systems.
How this is tested — this micro carries most of the 8- and 12-mark questions in the whole algorithms topic. It comes up two ways:
Paper 1 — structured question
- Part c: evaluate an automated decision system
- Part b: explain a concern about an algorithm in a scenario
Paper 2 — source-based question
- Q3: contrast two sources on an algorithmic decision
- Q4: synthesise sources, and Paper 3's intervention questions
The trap: listing harms: A dilemma has two sides that cannot both be satisfied. The efficiency that makes automated decisions worth having is the same property that makes them unreviewable one by one.
Discuss the use of an algorithm to decide which benefit claims are investigated for fraud.
Model answer plan
See the mark-by-mark plan — for / against / judgement, with marking guidance — in study mode.