A short history of artificial intelligence and what it can teach us

Transcript

1. Anissa Zacharias | November 7, 2025 A brief history of artificial intelligence and what it can teach us 2nd Workshop on Correctness and Reproducibility for Earth System Software
2. What this talk is not • Comprehensive • Exhaustive • Un-opinionated • Going to be concerned with anything else a philosophy professor would get excited about 2
3. What we’re talking about when we talk about AI 3
4. 4 History of AI as a field A non-exhaustive timeline 1950 1960 1970 1980 1990 2000 2010 2020 The Dartmouth Conference Perceptron ELIZA Jaberwacky A Statistical Approach to Language Translation IMB Deep Blue chess match IBM Watson Jeopardy appearance Siri OpenAI GPT-3
5. 5 History of AI significant criticism from the start 1950 1960 1970 1980 1990 2000 2010 2020 The Dartmouth Conference Perceptron ELIZA Jaberwacky A Statistical Approach to Language Translation IMB Deep Blue chess match IBM Watson Jeopardy appearance Siri OpenAI GPT-3 Lighthill Report Hubert Dreyfus Alchemy and Artificial Intelligence Hubert Dreyfus What Computers Can’t Do ❄
6. 6 History of AI significant criticism from the start 1950 1960 1970 1980 1990 2000 2010 2020 The Dartmouth Conference Perceptron ELIZA Jaberwacky A Statistical Approach to Language Translation IMB Deep Blue chess match IBM Watson Jeopardy appearance Siri OpenAI GPT-3 Hubert Dreyfus Alchemy and Artificial Intelligence Hubert Dreyfus What Computers Can’t Do “Most workers in AI research and in related fields confess to a pronounced feeling of disappointment in what has been achieved in the past twenty-five years. Workers entered the field around 1950, and even around 1960, with high hopes that are very far from having been realized in 1972. In no part of the field have the discoveries made so far produced the major impact that was then promised.” Lighthill Report
7. 7 History of AI significant criticism from the start 1950 1960 1970 1980 1990 2000 2010 2020 The Dartmouth Conference Perceptron ELIZA Jaberwacky A Statistical Approach to Language Translation IMB Deep Blue chess match IBM Watson Jeopardy appearance Siri OpenAI GPT-3 Lighthill Report Hubert Dreyfus What Computers Can’t Do Hubert Dreyfus Alchemy and Artificial Intelligence “It is the point of Dreyfus’s book that human and artificial intelligence are in fact quite different—in particular, that human intelligence is unique. Not only that, but a great mis- understanding accounts for public confusion about thinking machines, a misunderstanding perpetrated by the unrealistic claims researchers in AI have been making, claims that thinking machines are already here, or at any rate, just around the corner.” — Pamela McCorduck, Machines Who Think
8. Hubert Dreyfus What Computers Can’t Do Hubert Dreyfus Alchemy and Artificial Intelligence 8 History of AI significant criticism from the start 1950 1960 1970 1980 1990 2000 2010 2020 The Dartmouth Conference Perceptron ELIZA Jaberwacky A Statistical Approach to Language Translation IMB Deep Blue chess match IBM Watson Jeopardy appearance Siri OpenAI GPT-3 Lighthill Report Hubert Dreyfus What Computers Still Can’t Do “No piece of equipment makes sense by itself … What makes an object a chair is its function, and what makes possible its role as equipment for sitting is its place in a total practical context. This presupposes certain facts about human beings (fatigue, the ways the body bends), and a network of other culturally determined equipment (tables, floors, lamps), and skills (eating, writing, going to conferences, giving lectures, etc.)”
9. Hubert Dreyfus What Computers Still Can’t Do Lighthill Report 9 History of AI significant criticism from the start 1950 1960 1970 1980 1990 2000 2010 2020 The Dartmouth Conference ELIZA Jaberwacky A Statistical Approach to Language Translation IMB Deep Blue chess match IBM Watson Jeopardy appearance Siri OpenAI GPT-3 Hubert Dreyfus Alchemy and Artificial Intelligence Hubert Dreyfus What Computers Can’t Do Neural nets considered “too speculative” at the time to be truly AI and excluded from Computers and Thought, a book of influential AI papers. “AI is hard to define anyway; most of the time it ends up being viewed as those things w h i c h t h e A I p e o p l e a r e doing.” — Edward Feigenbaum Perceptron
10. Considerations for the AI of today 10
11. 11 products, generative AI, LLMs Considerations for the AI of today What we mean when we talk about AI
12. What we mean when we talk about AI 12 Considerations for the AI of today products, generative AI, LLMs machine learning, speech recognition, classification models, etc AGI
13. What we mean when we talk about AI 13 Considerations for the AI of today products, generative AI, LLMs machine learning, speech recognition, classification models, etc AGI hallucinations thinking
14. 14 Considerations for the AI of today Acknowledging limitations
15. 15 Considerations for the AI of today LLMs are not reproducible or verifiable Acknowledging limitations
16. 16 Considerations for the AI of today Objectivity Bias
17. Considerations for the AI of today 17 Inevitability
18. Sources and Supplemental Reading 18 github.com/anissa111/history-of-ai • IBM - The History of AI • Lighthill Report • Dreyfus - Alchemy and Artificial Intelligence • Dreyfus - What Computers Can't Do