Generative AI and Society
Lecturer: Prof. Dr. Orit Halpern (Office hour by appointment)
Courses time WinSe24/25
Week1: 2024.10.14 OSL3 (OpenScience Lab SLUB 12:00-16:00)
Week2: 2024.10.21 OSL3 (12:00-16:00)
Week3: 2024.10.28 OSL3 (12:00-16:00)
Week4: 2024.11.11 ABS/02-006 (August-Bebel-Str.20,30; 12:00-16:00)
Der Lehrraum befindet sich im sogenannten "blauen Haus", das ist das große Gebäude mit den blauen Elementen an der Fassade, quasi von der Straße aus links vom großen Hörsaal E011 (Kino im Kasten) aus gesehen.
Week5: 2024.11.25 ABS/02-006 (12:00-16:00)
Week6: 2025.12.02 OSL3 (12:00-16:00)
SYLLABUS
Welcome to Generative AI and Society
Today, we take it as almost natural that everywhere we go we can access “information” and interact with the environment. Shopping. Financial Transactions. Work. Are all mediated (in many places in the world) by machines. Increasingly, we assume ourselves (for better or worse) to be part of large, interconnected, and constantly evolving and learning networks. We expect our search engines to change their recommendations according to our searches, we assume that our consumer behaviors will be immediately responded to and fulfilled in ever faster and more accelerated manners, we expect fast fashion, fast downloads, and increased personalization. We have even come to believe that “smart” cities and grids, and AI infrastructures will help us better respond to catastrophes and manage large agglomerations of populations under rapidly changing security, economic, and environmental conditions. AI is seen as integral to our imagined extra-terrestrial future, and to our fantasies of a geoengineering that will save us from climate change. Whether we believe AI is the next step in “disaster capitalism” and extreme neo-liberalism and neo-colonialism, or whether we believe AI will finally set the market free and permit endless prosperity, health, and eternal growth, both points of view share in the faith that computing is ubiquitous, that the future of computation will no longer be about planning or representation, and that systems can autonomously evolve through constant reflexive analysis of data.
How did we come to believe these things? How has human survival come to be equated with increased computational penetration into the environment? How have technological and natural evolution become co-extensive? And how and why and for whom does the dream of governance without representation, or even programming (as in computer programming), serve?
This course will engage these questions. To get to better understanding of these issues, we will actually study many things that you may not consider “artificial” intelligence. We will examine histories of cybernetics, the social and human sciences, and the biological and natural sciences in order to understand how we came to see the computational management of the planet as a necessary and natural thing. The narrative of this course is that concepts of environment and ecology merged with those of computation and cognition to produce our contemporary discourse of AI. The course is roughly partitioned into three sections: Evolution, Intelligence, and Ecology. These categories are of course all contested, and quite slippery. But this is part of the challenge of this course. Collectively we will struggle to develop theories, narratives, and methods by which to engage contemporary discourses of AI, and to challenge them with alternative histories, vocabularies, and practices.
EVOLUTION
Week One:
What is Generative AI?
What are some of the stories we tell about generative AI? What is the Singularity? What is the concept of AI ascendance?
Please watch:
Roomba committing suicide: https://www.youtube.com/watch?v=GYFs3f7xf7c
Infinite Looping Siri, Alexa and Google Home: https://www.youtube.com/watch?v=vmINGWsyWX0
Geoffrey Hinton on BBC May 17, 2024
https://www.youtube.com/watch?v=MGJpR591oaM
Yuval Harari on AI/Daily Show
https://www.youtube.com/watch?v=euBAVec2RhE
Background on Neural Networks
Geoffrey Hinton Noble Prize Explanation on Physics (Popular physics explanation) (PDF)
Advanced Science Committee Nobel Prize (Geofrey Hinton Historical Explanation) (PDF)
Geoffrey Hinton Nature Deep review (PDF)
Optional
Yuval Harari on storytelling at Breuggruen Institute
https://www.youtube.com/watch?v=hXi-9fWASBY
Geoffrey Hinton on Nobel Prize
https://www.youtube.com/live/H7DgMFqrON0
All Watched Over by Machines of Loving Grace (Adam Curtis)
https://vimeo.com/groups/96331/videos/80799352
WEEK TWO: Evolution cont.
Yoshua Bengio et. al https://futureoflife.org/open-letter/pause-giant-ai-experiments/
The TESCREAL bundle: Eugenics and the promise of utopia through artificial general intelligence
Authors: Timnit Gebru, Émile P. Torres
DOI: https://doi.org/10.5210/fm.v29i4.13636 or https://firstmonday.org/ojs/index.php/fm/article/view/13636
Yann LeCann, The Path to General Artificail Intelligence: chrome-extension://efaidnbmnnnibpcajpcglclefindmkaj/https://openreview.net/pdf?id=BZ5a1r-kVsf
AND PLEASE GO BACK ON the article Geoffrey Hinton Nature Deep review (PDF)
If this is too much, PLEASE make sure to read the first two articles --gebru and bengio. The Le Cann is a little technical so just look at the first few paragraphs and the conclusion to understand it.
ASSIGNMENT:
Please come in with one-two paragraphs on your portolio/ possible final project for the course. THis should be an independent research/app/ multi media project you wish to do. If you do this for the portoflio its not expected to be finished this semester, we will set a milestone. If its just for the class, the expectation is a 15 page equivalent (I do accept other formats in discussion with myself) research paper or creative project equivalent.
Second assignment, please prepare three questions about the readings to be asked in class, or three things you wish to discuss. Also be prepared to do some prompt playing in class, that we will analyze.
Optional Readings:
Stephen Gould, The Mismeasure of Man. New York: Norton (1996) pp. 51-104
Simone Brown, Dark Matters, (Durham: Duke University Press, 2015)
Introduction, Chapter 4, Epilogue.
Darwin, Origin of the Species, (1859) http://darwin-online.org.uk/converted/pdf/1861_OriginNY_F382.pdf
Introduction/ Chapter IV-Survival of the Fittest/Chapter V: Laws of Variation, Chapter XV--Recapitulation and Conclusion.
Goodfellow, Ian, et al. "Generative Adversarial Networks." Association for Computing Machinery (ACM): Association for Computing Machinery (ACM), 2020. Print. (PDF)
INTELLIGENCE?
Week Three: Intelligence
Lorraine Daston, Rules and Rationality: How Reason become Rationality (docx online)
Orit Halpern, Financializing Intelligence, E-flux Architecture, March 2023
Asjosj, Vaswani., et al. "Attention Is All You Need." 31st Conference on Neural Information Processing Systems. 2017. Print. (PDF)
Week Four: Whose Intelligence? What Species? What Politics
Louise Amoore a, Alexander Campolo a, Benjamin Jacobsen b, Ludovico Rella a,”A world model: On the political logics of generative AI”, Political Geography
Volume 113, August 2024, 103134 https://durham-repository.worktribe.com/output/2467333/a-world-model-on-the-political-logics-of-generative-ai
Zuboff, Shoshana. The Age of Surveillance Capitalism. New York: Hachette Book Group, 2019.(Excerpts) https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2594754
Steven Johnson, “Looking into the Future has always defined Humanity, Will AI be the best crystal ball of all”, NYT, special issue Tech and Design Issue, November 15, 2019: https://www.nytimes.com/interactive/2018/11/15/magazine/tech-design-ai-prediction.html?rref=collection%2Ftimestopic%2FArtificial%20Intelligence&action=click&contentCollection=timestopics®ion=stream&module=stream_unit&version=search&contentPlacement=1&pgtype=collection (You might want to look at the whole issue).
Carl Zimmer, “Yes, the Octopus Is Smart as Heck. But Why?” New York Times, Nov. 30, 2018
For assignments:
1) Please submit the page and half on your project (final) if you have not done so, please also submit a bibliography.
2) Please come prepared with a summary of at least two articles, three questions you want to discuss, and a prepared statement if called upon as to why you think this is about AI.
3) Please find one news paper clipping or post or something that discusses what constitutes humanity in our present? or explores ideas of intelligence in animals o rmachines but with a focus on what is intelligence.
ENVIRONMENT
Week Five: Environment as Artificial Intelligence
Halpern, Orit. "Planetary Intelligence." The Cultural Life of Machine Learning. Ed. Roberge, J., Castelle, M. London: Palgrave Macmillan, 2021. 227-56. Print. (PDF)
Jennifer Gabrys: (2016). Program Earth: Environmental Sensing Technology and the Making of a Computational Planet. Minneapolis: University of Minnesota Press.1-40
--Becoming Planetary, https://www.e-flux.com/architecture/accumulation/217051/becoming-planetary/
Haff, Peter K. "Technology as a Geological Phenomenon: Implications for Human Well-Being." A Stratigraphical Basis for the Anthropocene. Eds. Waters, C. N., et al. Vol. 395. London: Geological Society, 2014. 301–09. Print. (PDF)
WEEK SIX: THE FUTURE OF INTELLIGENCE
Vinge, Vernor. "Technological Singularity." (1993) (PDF)
Ray Kurzweil, The Singularity Is Near (just read intro by Kurzweil) (2005) (PDF) pp. 21-44,44-56,Epilogue
Yann Le Cunn: Path Towards Autonomous AI (PDF)
ASSIGNMENT:
Over the two weeks in November when we do not meet you are to read the Montreal Declaration of Artificial Intelligence
https://montrealdeclaration-responsibleai.com/the-declaration/
and the EU AI Act Summary:
https://artificialintelligenceact.eu/the-act/
Together as a class we will write our own declaration for responsible development of Artificial Intelligence and our own Act for what we think might be a better future way to govern AI. Each student will pick one part of one of these documents to re-imainge, or you can work in groups. The final assignment will be your own creative portion of an AI Ethics Declaration. You can re-write on portion of the law, such as requirements for High Risk, or one element of the declaration, you need to imagine this document and situate what you think might be a better type of ethics. Perhaps one that includes more forms of intelligence, produces different ideas of rationality and optimization, or some other concept.
OR
You can produce a dictionary of new terms you want applied to Artificial Intelligence. You should propose 4-5 terms that might be new and unused in the literature you have read. For example, replacing ideas of optimization, efficiency, etc.
Each assignment can be done alone or in a group. If alone you should hand in the equivalent of 10 pages double spaced 12 point font of something Like Times New Roman, 1 inch margins. And if you are in groups, depending on the size of the group the project should be 25-30 pages, similar specs.
These projects will be due on March 1, 2025.