The guide names fairness and bias, accountability, transparency, uneven laws and regulation, and automation displacing people. Anchoring all of them to one real decision is what turns a list into an argument.
One job application scored by a model, and each dilemma it runs into.
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 that makes the topic genuinely hard, and it is what the COMPAS argument was actually about.
One fairness test
- The system is right about the same share of each group
- It passes
- The company used this test
Another fairness test
- When it is wrong, it is wrong in the same direction for each group
- It fails
- The journalists used this test
Both were right: Except in special cases the two tests cannot both be satisfied. That is a mathematical result, not a dispute about facts — so “make it fair” is not an instruction a developer can simply follow.
See how examiners mark answers
Access past paper questions with model answers. Learn exactly what earns marks and what doesn't.
The guide names uneven and underdeveloped laws, regulations and governance. That is changing, and knowing roughly where it has got to is worth a mark and a half.
Real-world examples you can name
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.
The Bletchley Declaration on AI safety — November 2023
The first international statement on frontier AI risk signed by both the US and China, agreeing that the most capable models pose risks worth coordinating on. It is a declaration of intent, with no enforcement.
Who it affected: Governments, model developers, and anyone hoping governance keeps pace.
San Francisco's facial recognition ban — May 2019
The first major US city to bar its own agencies, including the police, from using facial recognition — a rule about government use, not about the technology existing. Several other cities followed, and a few later relaxed theirs.
Who it affected: Residents, and police forces that had to find other methods.
The pattern to notice: Regulation so far is mostly about use, not about the technology: who may deploy it, for what, with what duties. That is the shape any intervention you propose should take.
How this is tested — this micro carries most of the long questions about AI, at both levels. It comes up two ways:
Paper 1 — structured question
- Part c: evaluate an AI system making decisions about people
- Part b: explain a specific concern about a described system
Paper 2 — source-based question
- Q3: contrast two sources on an AI controversy
- Q4: synthesise sources, and the HL Paper 3 intervention
The trap: writing about robots taking over: Real questions are about narrow systems doing what they were built for, and getting it wrong in ways nobody can see.
Evaluate the use of artificial intelligence to decide which patients are seen first in a hospital.
Model answer plan
See the mark-by-mark plan — for / against / judgement, with marking guidance — in study mode.