A Short History of Artificial Intelligence (TU Vienna, Winter 2024)

A final writing project will take the place of an exam. Discussion participation and smaller writing assignments will also contribute to the grade. These elements are weighted as follows:

• 2 short papers, 15%x2 = 30%. Due November 17 and December 3.
• discussion participation = 35%
• project (due end of semester exact date TBA) 35%

Attendance and Participation Requirements

My basic expectation for full credit for course participation would be that each person make two significant contributions to discussion in each course meeting. These should be relevant to the question asked and show knowledge of the course readings. Sharing opinions and experiences is also encouraged, but cannot substitute fully for showing knowledge of the readings. So do not feel that you have to dominate the discussion to score well – there is enough time in the course meeting for all students to earn full credit.

Attendance is required, and missing class will hurt your grade or even prevent you from completing the course. Specifically:

1. Students can miss one class without penalty. The first missed class will be treated as if you came but did not speak.
2. Students who miss more than one class without making up for the absence will be penalized in the calculation of their grade (when I average the component scores and apply thresholds to set an overall grade).
3. Making up a class involves emailing me several paragraphs of text, total around 400 words, as an informal essay that takes an element of the assigned chapter from my book and connects it to material from at least one of the other readings for the missed class. As long as your responses prove you did the reading, I encourage you to include personal ideas and opinions rather than just summarize chunks of the text. (Also, no AI generated summaries!)
4. If you enjoy the class but do not feel comfortable talking, you can do the same thing even for classes you do attend to improve your participation score.
5. Any student with four or more unexcused absences will be graded as not having completed the course.

Class Meeting Schedule and Readings

All readings have been provided to enrolled students via a Dropbox folder. Please read the indicated documents BEFORE each course meeting. The core content for each session is the chapter from my preprint book, working title Artificial Intelligence: The History of a Brand. We will be reading one chapter for each course meeting. The other readings are chosen to complement that chapter. As each chapter is split between general narrative and a case study of a specific historically important AI system, the readings are split between information on the case study system and historical sources that go with the general story. In historical terms, the readings are a mix of primary sources (those from the time period concerned) and secondary sources (more recent research papers by historians studying these topics). So you will see a lot of variety in length, tone, etc. in the readings

1. Nov 5: Introduction: The Brand That Wouldn’t Die
   
   1. "The Intelligence Age." samaltman.com, 2024, https://ia.samaltman.com/.
   2. Simon, Herbert A. "The Corporation: Will It Be Managed by Machines?". In Management and Corporations 1985, edited by Melvin Anshen and George Leland Bach, 17-55. New York: The McGraw-Hill Book Company, 1960. Selected pages only.
   3. Darrach, Brad. "Meet Shakey: The First Electronic Person." Life, November 20, 1970, 58b onward.
   4. Raphael, Bertram. The Thinking Computer: Mind inside Matter. San Francisco, CA: W. H. Freeman & Company, 1976 Epilogue only.
   5. Feigenbaum, Edward A, Pamela McCorduck, and Penny Nii. The Rise of the Expert Company: How Visionary Companies Are Using Artificial Intelligence to Achieve Higher Productivity and Profits. New York: Times Books, 1988. Chapter 1 only.
2. Nov 7: The Birth of a Brand (General Problem Solver), 1950s
   
   1. "A Proposal for the Dartmouth Summer Research Project on Artificial Intelligence." Professor John McCarthy: Father of AI, Updated August 31, 1955, http://jmc.stanford.edu/articles/dartmouth/dartmouth.pdf.
   2. Dick, Stephanie. "Of Models and Machines: Implementing Bounded Rationality." Isis 106, no. 3 (2015): 623-34.
   3. Slagel, James R. Artificial Intelligence: The Heuristic Programming Approach. New York: McGraw-Hill, 1971. Chapter 8 (“The General Problem Solver”) only.
3. Nov 11: Institutionalizing the AI Brand (SHRDLU),1960-75
   
   1. Penn, Jonnie. "Animo Nullius:: On Ai’s Origin Story and a Data Colonial Doctrine of Discovery." BJHS Themes 8 (2023): 19-34.
   2. Elzway, Salem. "Armed Algorithms: Hacking the Real World in Cold War America." Osiris 38 (2023): 147-64.
   3. Winston, Patrick. Artificial Intelligence. Reading, MA: Addison-Wesley, 1977. Chapter 6 only.
   4. Youtube video of scripted demo https://www.youtube.com/watch?v=bo4RvYJYOzI
   5. Optional bonus reading: Winograd, Terry. Oral History Interview by Arthur L. Norberg. . Minneapolis, MN: Charles Babbage Institute, 1991.
4. Nov 13: Challenges to the AI Brand (Hearsay II), 1965-1980
   
   1. Dreyfus, Hubert L. Alchemy and Artificial Intelligence. Santa Monica, CA: RAND Corporation, 1965. (This is quite long, so focus on pages 2-17 and 75-86 for the introduction and conclusions).
   2. Agar, Jon. "What Is Science For? The Lighthill Report on Artificial Intelligence Reinterpreted." British Journal for the History of Science 53, no. 3 (September 2020): 289-310.
   3. Erman, Lee D, Frederick Hayes-Roth, Victor R Lesser, and D Raj  Reddy. "The Hearsay-Ii Speech-Understanding System: Integrating Knowledge to Resolve Uncertainty." Computing Surveys 12, no. 2 (June 1980): 213-53.
   4. Wilks, Yorick. "Time Flies Like an Arrow." New Scientist, December 15, 1977, 696-98.
5. Nov 14: Branding with Knowledge (Mycin), 1975-1985 (end week 2)
   
   1. Schank, Roger C, and Robert F Abelson. "Scripts, Plans, and Knowledge." In Proceedings of the Fourth International Joint Conference on Artificial Intelligence, edited by International Joint Conference on Artificial Intelligence, 151-57. San Francisco, CA: Morgan Kaufmann, 1975.
   2. November, Joseph. Biomedical Computing: Digitizing Life in the United States. Baltimore, MD: Johns Hopkins, 2012. Pages 242-268 only.
   3. Buchanan, Bruce G, and Edward Hance Shortliffe, eds. Rule-Based Expert Systems : The Mycin Experiments of the Stanford Heuristic Programming Project. Reading, MA: Addison-Wesley, 1984. Chapters 1, 15 & 30 only.
6. Nov 18: Selling AI (Cyc), 1980s
   
   1. Anonymous. "Artifical Intelligence: The Second Computer Age Begins." Business Week, March 8, 1982, 66-75.
   2. Barker, Virginia E, and Dennis E O'Connor. "Expert Systems for Configration at Digital: Xcon and Beyond." Communications of the ACM 32, no. 3 (March 1989): 298-318.
   3. McDermott, Drew, M Mitchell Waldrop, Roger Schank, B. Chandrasekaran, and John McDermott. "The Dark Ages of Ai: A Panel Discussion at Aaai-84." AI Magazine 6, no. 3 (1985): 122-34.
   4. Schank, Roger C. "Where's the Ai?". AI Magazine 12, no. 4 (1991): 38-49.
   5. Ekbia, Hamid R. Artificial Dreams: The Quest for Non-Biological Intelligence. New York: Cambridge University Press, 2008. Chapter 4 only.
7. Nov 20: Out of Fashion, AI Tries New Things (Dragon Naturally Speaking), 1990s
   
   1. Brooks, Rodney A. "Intelligence without Representation." Artificial Intelligence 47 (1991): 139-59.
   2. Russell, Stuard J. Profile of Judea Pearl, ACM A.M Turing Award Website, https://amturing.acm.org/award_winners/pearl_2658896.cfm.
   3. Li, Xiaochang. "'There’s No Data Like More Data': Automatic Speech Recognition and the Making of Algorithmic Culture." Osiris 38 (2023): 165-82.
   4. Jones, Matthew L. "How We Became Instrumentalists (Again): Data Positivism since World War II" Historical Studies in the Natural Sciences 48, no. 5 (November 2018): 673-84.
8. Nov 22: Machine Learning Becomes the Hot New Brand (AlexNet), 1980s-2012 
   
   1. Mendon-Plasek, Aaron. "Irreducible Worlds of Inexhaustible Meaning: Early 1950s Machine Learning as Subjective Decision Making, Creative Imagining and Remedy for the Unforeseen." BJHS Themes 8 (2023): 65-80.
   2. Law, Harry. "Bell Labs and the 'Neural' Network." BJHS Themes 8 (2023): 143-54.
   3. Krizhevsky, Alex, Ilya Sutskever, and Geoffrey E. Hinton. "Imagenet Classification with Deep Convolutional Neural Networks." Communications of the ACM 60, no. 6 (June 2017): 84-90. (Reprint of 2012 conference paper)
   4. Metz, Cade. Genius Makers: The Mavericks Who Brought Ai to Google, Facebook, and the World. New York: Dutton, 2021.
9. Nov 25: Thanks to Chatbots, AI Finally Conquers the World (ChatGPT), Present day
   
   1. Bender, Emily M., Timnit Gebru, Angelina McMillan-Major, and Margaret Mitchell. "On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?". In Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency (Facct '21), 610-23. New York: Association for Computing Machinery, 2021.
   2. Vinge, Vernor. "Technological Singularity." Whole Earth Review, Winter, 1993, 88-95.
   3. Weise, Karen, Cade Metz, Nico Grant, and Mike Isaac. "Inside the A.I. Arms Race That Changed Silicon Valley Forever." December 5 2023. https://www.nytimes.com/2023/12/05/technology/ai-chatgpt-google-meta.html.
   4. Hermann, John. "Sam Altman and Open Ai Are Victims of Their Own Hype." Intelligencer Updated November 22, 2023, https://nymag.com/intelligencer/2023/11/how-big-techs-ai-hype-cycle-swallowed-sam-altman-openai.html.
   5. "OpenAI Takes Its Mask Off." OpenAI Takes Its Mask Off, The Atlantic, Updated September 24, 2024, https://www.theatlantic.com/technology/archive/2024/09/sam-altman-openai-for-profit/680031/.
10. Nov 27: What Was AI, Anyway?

Short Papers

Assignment 1 (Due Sunday, November 17 2024)

Covers material from course meetings 1-5 which follow the story of AI from 1955 through to the late-1970s. Answer ONE of the following questions:

1. The Dartmouth proposal of 1955 set out some ambitious goals. Over its first two decades, did AI researches succeed in making significant progress toward them?
2. Most modern AI relies on entirely different technical approaches to those that dominated the field up to the 1990s. Does that mean that the history of 20^th century AI is irrelevant to present day AI specialists?
3. While early AI researchers tended to see their work as non-political, historians often focus on the connections of their labs to the US military, capitalism, and even colonialism. Was AI corrupted, or at least shaped, by these broader social factors?

Assignment 2 (Due December 3, 2024)

Covers material from the entire course, but you should focus more on readings from the second half.

1. Judea Pearl has complained that the current focus on big data approaches to AI means that vast sums of money are being spent "building cathedral to a handcuffed god" while researchers taking other approaches are deprived of support. Why do Pearl, and other advocates for symbolic AI, feel that neural nets and other big data technologies do not on their own provide a foundation for true intelligence?
2. Back in the 1970s, neural nets seemed like a failed old technology. By the 1990s they were fashionable again, and now they underpin the biggest ever AI boom. What caused this remarkable comeback?
3. In the 1970s and 80s, AI researchers and textbooks mostly avoided discussion of the Turing Test and the possibility of superhuman intelligence. Today leading researchers and the heads of AI companies debate when AGI will arrive and the odds of human extinction as a result. What caused this change?

Guidelines (Applies to both of the short paper assignments)

• Your paper should be between 1,000 and 1,200 words.
• Make sure that you explicitly answer the assigned question, which means choosing a side “yes” or “no” and explaining why in a thesis statement at the start of the paper. Then use the rest of the paper to support your answer with detail and evidence.
• Support your answer with specific facts from the assigned readings. These assignments are your main chances to show that you have carefully done the readings and thought enough about them to start making connections between them. So make sure you references at least five readings from three different class meetings. (In this context, you can count each chapter of my book as a different reading).
• Using quotes from the readings is good, but do not quote entire paragraphs. Mix short quotes with summary, to make your own points, the way I do in the book.
• Cite your sources, even those assigned for class. The course site includes full citations for all assigned readings in Chicago format, but you can use whatever your favorite citation system is.
• Supplementing the assigned readings with other relevant sources such as technical papers from this era (primary source) or work by historians (secondary sources) is encouraged but not required. This cannot substitute for reference to the assigned readings.
• I know that getting an AI to write your AI history paper is conceptually provocative, but please don’t.
• Submit your paper via the submission link I will provide. This will be private -- other students will not be able to see what you submit.