COMP90087 · The Ethics of Artificial Intelligence
The History of Artificial Intelligence
Week 1 opens COMP90087 by tracing AI from Turing and symbolic reasoning through two AI winters to the machine-learning revival, and by drawing the ethics-versus-law distinction that recurs all semester. In the closed-book hurdle exam this surfaces as definitional MCQs — most reliably “what is an AI winter?” — and the diversity/hype themes seed short-answer and essay material. It also frames the whole subject’s thesis: powerful tools carry responsibility, so “just because you can does not mean you should.”
What this chapter covers
- 01Ethics ≠ law: law = what the state deems legal/illegal; ethics (moral philosophy) = what is right/wrong, good/bad character
- 02Three reasons to study AI history — intrinsic interest, hype (repeated failed predictions), and diversity of who builds AI
- 03Computing milestones sequence: Antikythera (≈87 BCE, not programmable) → Pascaline → Jacquard loom (punch cards) → Babbage’s Analytical Engine (Ada Lovelace = first programmer) → ENIAC → CSIRAC
- 04The Dartmouth Workshop (1956) coined “artificial intelligence”; the divide-and-conquer subfields (perception, knowledge representation, learning, reasoning, planning)
- 05Golden Age (reasoning as search: A*, STRIPS, the perceptron) → First AI winter → Knowledge Era (expert systems, MYCIN) → Second AI winter → Revival (compute, big data, backpropagation)
- 06AI winter (exam definition): a period of drastically reduced funding, interest and progress caused by hyped expectations going unmet
- 07Causes of the winters: scalability, commonsense knowledge, perceptron limitations, Moravec’s paradox, the qualification problem, maintenance
- 08Amara’s Law and diversity/representation (Lovelace, Turing, Hopper, Fei-Fei Li, Gebru) as sources of present-day ethical gaps
Exam-style short answer: define the “AI winter” and diagnose a cause
- +1Mark 1 — definition. An AI winter is a period of drastically reduced funding, interest and progress in AI research, triggered when hyped expectations are not met. Do not answer “a cold season” or “a ban on AI” — those are the distractor traps.
- +1Mark 2 — a cause of the first winter (1974–1980). Name any one genuine cause: scalability (everything had to be hand-programmed), the difficulty of encoding commonsense knowledge, the perceptron’s limitations (Minsky’s critique), or Moravec’s paradox (child-level perception is harder for machines than “high-level” tasks like chess).
- +1Mark 3 — ethical relevance. History matters because grand, repeatedly-failed predictions show that uncritical hype can be irresponsible — Amara’s Law (we overestimate a technology’s short-run effect and underestimate its long-run effect) is the frame examiners reward.
Key terms
- AI winter
- A period of drastically reduced funding, interest and progress in AI research, caused by hyped expectations not being met. Two are named in the subject: 1974–1980 and 1987–1993.
- Ethics vs law
- Law is what the state deems legal or illegal; ethics (moral philosophy) is about what is right or wrong, good or bad, and good or bad character. Related but distinct — something can be legal yet unethical.
- Dartmouth Workshop (1956)
- The event widely treated as the “birth” of AI, where the term “artificial intelligence” was coined and AI was split into subfields: perception, knowledge representation, learning, reasoning and planning.
- Moravec’s paradox
- The observation that tasks humans find easy (perception, navigation) are hard for machines, while “high-level” tasks like chess are comparatively easy — a cause cited for both AI winters.
- Amara’s Law
- “We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.” The standard frame for reasoning responsibly about AI hype.
- Expert system
- A knowledge-era architecture (user interface + inference engine + knowledge base) that encodes an expert’s rules; MYCIN diagnosed blood diseases sometimes better than doctors but made mistakes and was not trusted.
The History of Artificial Intelligence FAQ
Do I need to memorise every date and machine in the AI-history timeline?
Not every date, but you should know the sequence and what each milestone was or was not — especially “programmable” vs “general-purpose” (the Jacquard loom is programmable but not a general computer; Babbage’s Analytical Engine is the general-purpose design, with Ada Lovelace as first programmer). The highest-value single item is the AI-winter definition, followed by the two winters’ causes and the revival triggers (compute, big data, backpropagation).
What is the difference between ethics and law, and why does Week 1 stress it?
Law is what the state makes legal or illegal; ethics is about right and wrong and good or bad character. The subject stresses it because much of AI is currently legal yet ethically contested — the whole point of the course is to reason about what we ought to do, not only what we are permitted to do. Expect it as a clean MCQ distinction.
Why does the subject talk so much about diversity in AI history?
Because a lack of diversity among AI’s builders is presented as a root cause of present-day ethical gaps in privacy, safety, fairness, transparency and accessibility — design decisions, data and attitudes reflect who makes them. The under-represented contributors highlighted (Lovelace, Turing, Grace Hopper, Fei-Fei Li, Timnit Gebru) give you concrete examples for a short answer or essay.
Can AI help me revise the history-of-AI material for COMP90087?
Yes, as a study aid. Sia is an AI tutor built to mirror how COMP90087 is taught at the University of Melbourne: it can quiz you on the milestone sequence, drill the AI-winter definition and its causes, and check your ethics-versus-law answers step by step. It explains the reasoning and does not complete graded assessment for you — University of Melbourne academic-integrity rules apply, so use it to understand the material, not to generate work you submit.
Exam move
Build a one-page timeline you can redraw from memory: milestones → Dartmouth (1956) → Golden Age → first winter → Knowledge Era → second winter → revival, with one cause tagged to each winter and one enabler tagged to the revival. Lock the AI-winter definition verbatim (funding + interest + progress collapsing) because it is the most reliable MCQ, and pair Amara’s Law with two failed predictions so you can also handle the hype theme in a short answer. Keep a short list of under-represented contributors for the diversity essay angle. This is examinable in the closed-book hurdle exam, so drill it into recall now rather than leaving it to SWOTVAC.
Working through The History of Artificial Intelligence in COMP90087? Sia is AskSia’s AI AI Ethics tutor — ask any COMP90087 The History of Artificial Intelligence question and get a clear, step-by-step explanation grounded in how COMP90087 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.
