Showing posts with label Social Network Analysis. Show all posts
Showing posts with label Social Network Analysis. Show all posts

Sunday, January 5, 2025

Network Ghosts in the Age of Generative AI

 

What are faculty thinking about generative AI? In my role at our faculty center, I speak to faculty often on the problems they face teaching in the era of AI, and the workarounds they've come up with. The advent of publicly available generative AI platforms was not something people in my field (anthropology) or other faculty in the social sciences and humanities were clamoring for. And yet here we are. This has led to many responses: anguish, certainly, but also ways of incorporating--or at east channeling--the usage of generative AI in the classroom.

But what about faculty outside of my university? I used NodeXL to download Reddit data from the "/Professors" subreddit using the keyword "AI." This generated records of about 2500 users posting, commenting or replying for a total of 7000 contributions to the debate. I then grouped the data in clusters of similar postings, and abstracted the top words from each group as indicated by "up-vote" (which functions as more of a "like" in Reddit). As you can see, faculty were not particularly optimistic about AI in 2024. Yes, there were a couple of more computopian posters (and at least one computer scientist) who chided the community for rejecting what they saw as inevitable. But most worried that their efforts to teach writing, critical thinking, methodology and analysis were thwarted by student reliance on generative AI. Cynically, they predicted their university's tolerance for AI cheating, and speculated over their ability to continue as faculty under these conditions.

In 2024, Reddit sold their content to Google to train their large language models. This would have been been more objectionable, perhaps, if it wasn't already abundantly clear that generative AI have already been trained on Reddit, which maintains a relatively open API at a time when most social media have monetized their social network data. But what happens to that Reddit data when its re-constituted by generative AI? I decided to prompt Microsoft's Co-Pilot (to which I have enterprise-level access) to generate a spreadsheet of a Reddit conversation on AI between professors. Here's the prompt: "I would like you to generate an excel file similar to a Reddit conversation on a subreddit called "professors." The posts should discuss ChatGPT and student work from the perspective of the professor, and should include comments and replies to those comments. There should be 4 columns in the spreadsheet: A (person commenting or replying); B (person whom A is replying to); C (the text of the comment or reply); and D (the date of the reply or comment). Please populate the spreadsheet with at least 20 comments and 350 replies to those comments."

Co-pilot returned a network with with just 10 users, with 350 edges representing multiple re-postings(?) of user posts. Re-posting really isn't a thing with Reddit, so perhaps there's some confusion here with XTwitter. Since this is a much smaller network, I just labeled the 10 nodes with key words from their posts. The comments are a near "upside down" to the actual Reddit discourse over 2024, generally praising the efficiencies of generative AI and, when critical, speculating over the need for faculty at all (hence the precarity). Of course, there's a snarky comment on "Clippy," the irritating Microsoft assistant. The network itself, while smaller, is also structurally different. The actual Reddit network has a density of .001158737. In network measures of density, "1" would represent 100% connection--everyone connected to everyone else. So .0012 may not seem like much, but it's typical of social media networks where, after all, most of us don't feed the trolls and we save our replies for issues (and users) that we really care about. On the other hand, my AI-generated network has one of 0.966666667--an almost perfectly connected network where everyone has replied to everyone in a style of a polite and ploddingly inclusive panel discussion.




So, I guess that Co-Pilot does a lousy job simulating a subreddit? Yes, but, I think, more than that. It wasn't that long ago (2023) when XTwitter adopted a fee-based model for API access. That decision placed Twitter data beyond the reach of most of us. When social media data disappears behind paywalls, we (ordinary researchers) no longer really have access to the "connected action" of social media. While we can certainly look at social media, this only exposes us to our respective corners of the media platforms we inhabit and the structural components of social media are lost. But what happens when social media content is sold to OpenAI or Google Gemini? When social media disappears into a large language model, both the content and the connections are lost, and the simulated networks produced through generative AI manage to misrepresent social media on both fronts. Since Co-Pilot's inner workings are opaque to us, it is unclear if these results are the result of deliberate choice, unintended bias or something else.

Thursday, October 3, 2019

Futures at AAACASCA 2019

The American Anthropological Association/ Canadian Anthropological Society meeting in Vancouver is in November, but the browsable schedule is already out.  As in previous years, I have identified future-oriented or science fiction-oriented panels that I would love to attend (including two I'm on). This may not be a complete list, and I apologize for panels I've missed.  But even this, incomplete as it might be, is an impressive collection of a robust future-orientation in the work of anthropologists. 



Thursday, November 21

8:00 AM – 9:45 AM  –  Decolonial Belongings and Futures: Creating Spaces of Belonging thru Epistemic Disobedience - Vancouver CC EAST, Room 7
2:00 PM – 3:45 PM  –  Biofutures - Vancouver CC WEST, Room 122
2:00 PM – 3:45 PM  –  Haunting Toward the Future: Colonial Durabilities and Temporalities - Vancouver CC EAST, Room 13
2:00 PM – 3:45 PM  –  Untaming futures? Plural knowledges, unknown environments and technologies of anticipation (Part 1) - Vancouver CC WEST, Room 202
4:15 PM – 6:00 PM  –  In an Atmosphere of Change: Speculative Futures in Anthropological Perspective - Vancouver CC WEST, Room 118
4:15 PM – 6:00 PM  –  NARRATING THE FUTURE FOR A WARMING WORLD - Vancouver CC WEST, Room 205
4:15 PM – 6:00 PM  –  Untaming futures? Plural knowledges, unknown environments and technologies of anticipation (Part 2) - Vancouver CC WEST, Room 202

Friday, November 22

10:15 AM – 12:00 PM  –  Utopia and Changing the Future: Anthropology’s Role in Imagining Alternatives (Part 1) - Vancouver CC EAST, Room 11

2:00 PM – 3:45 PM  –  The Climate of Governance and the Governance of Climate: Negotiating the Futures of Natures & Cultures - Vancouver CC EAST, Room 15
2:00 PM – 3:45 PM  –  Utopia and Changing the Future: Anthropology’s Role in Imagining Alternatives (Part 2) - Vancouver CC EAST, Room 11
4:15 PM – 6:00 PM  –  Horizons of Possibility: Dynamic Future Selves in a Changing and Contested World - Vancouver CC WEST, Room 115

Saturday, November 23

8:00 AM – 9:45 AM  –  Algorithmic Futures: Computing as a Site and Object of Technopolitical Interventions - Vancouver CC WEST, Room 301
8:00 AM – 9:45 AM  –  Forging Futures in Contested Landscapes - Vancouver CC WEST, Room 215
10:15 AM – 12:00 PM  –  Forecasting Futures: Education as Speculative Practice - Vancouver CC WEST, Room 119
10:15 AM – 12:00 PM  –  So many futures, so little time: Anthropological approaches to catastrophe and the future - Vancouver CC EAST, Ballroom C
4:15 PM – 6:00 PM  –  Ethnographies of Palestinian Futures - Vancouver CC WEST, Room 204

Sunday, November 24

10:15 AM – 12:00 PM  –  Geological Anthropology: Waters, Ruins, Futures (Part 2) - Vancouver CC WEST, Room 101 & 102

The Meaning of the Future

Yet there is a great deal of polysemy implied in "the future," and our orientation to future temporalities likewise varies (Bryant and Knight 2019).  I did some text analysis of the abstracts for these panels in order to look at the evolving terrain of future work [click on the graph for the full size]:

 



The graph uses "Infranodus,"a web-based, text analysis application that uses word co-occurence to construct a network.  Nodes are key terms, and the edges (or lines) between them show words (actually lemmas) separated by 1 word or words separated by two words (Paranyushkin 2019).

Additionally, the algorithm tries to identify "clusters" of terms--represented by different colored nodes and edges here.  But this seems of limited efficacy here, where there is considerable overlap in the nomenclature of the future.  Better, perhaps, is to focus on a few key terms, and the terms to which they're linked.

"Environmental" [click on the map for a full image of the network]



"World"


"Climate"


"Alternative"


These keywords, together with the connections they forge, ultimately tell a more nuanced story about anthropology's emerging futures.  The lemma "world" might appear in texts as "worlding," "world-building" etc., and might point, on the one hand, to the changes inextricably impacting our world today.  On the other hand, "world" also includes links to the prospect of different worlds, however defined, whether in "space" or "imagined."  "Alternative" opens on to the imaginative element of anthropological futuring, and the ways this might gesture towards other possibilities less premised on capitalist exploitation.  This includes indigenous futures, and alternative narratives on the future from oppressed peoples.  On the other hand, "climate" brings us into the decidedly more pessimistic futures of the anthropocene, where "change," "health" and "environment" make up the dreadful calculus of environmental catastrophe.

All together, the pessimism and the optimism of the present moment, one where we teeter on the brink of future disaster, while alternatives appear to us (as anthropologists) in multiple forms, from policy changes, to space travel, to worlds re-shaped by alternatives to Eurocentric capitalist exploitation.  The future work evolving in anthropology engages all of these levels simultaneously: 1) the future as a significant horizon in the lives of our interlocutors; 2) the future as an ethnographic object in its own right; 3) the future as a site for anthropological interventions.

References

Bryant, Rebecca and Daniel Knight (2019).  The Anthropology of the Future.  NY: Cambridge University Press.

Paranyushkin, Dmitry (2019).  "Infranodus."  In Proceedings of WWW '19: The Web Conference (WWW '19), May 23, 2019, San Francissco, CA.  


Saturday, October 20, 2018

Mapping the Future at the American Anthropological Association Annual Meeting

"The future" (however imagined) continues to be a concern for anthropologists, and this year is no different than 2017.  But while I was content to just list the different panels in 2017, this year I decided to construct a semantic map of the session abstracts.  First, I created a text document with each of the 28 session abstracts that explicitly concerned the future as an object of research (rather than something like "the future of graduate education").  Then, I loaded up the file into Cowo, which spit out 55 words by frequency of occurrence (minus all of the stop words like "the").  Then I loaded the file onto VOSviewer, and created a semantic map of co-occurrences between terms (nodes) in the same sentences.

Here's the visualization from VOSviewer:

  
And here it is again in Gephi:

We can identify several semantic clusters here, but I want to highlight a few: 1). urban resistance to the neoliberal (right); 2) environmental disaster and the future of the anthropocene (bottom); 3). the utopian imagination for critical alternatives (left); and 4). human migration and human futures in an age of increasing precarity (top). 

This semantic map is a a helpful shorthand for taking the pulse of the future in anthropology right now.  Hurtling toward disaster along multiple axes simultaneously (environmental, political, demographic), anthropologists (and their interlocutors) occupy multiples sites of emergence across precarious futures. 

Are there themes that bring together these different future orientations?  Here are the top terms as defined by betweenness centrality:


era
ecological
anthropologist
city
infrastructural
mobility
migrant
precarious
urbanization
ethnographic
anthropology
emergent
institutional
cultural
modernity
politically
anthropological
power
roundtable
utopian
film  

This AAA promises to consider futures that impinge onto anthropological presents--that is, ecological and urban catastrophe that emerges into coeval fieldsites.  Yes - there's still a concern here for utopian promise (there's a panel on Ursula K. Le Guin!), but much of the panels in this map consider the disastrous coincidence of precarious futures with precarious presents. 

Thursday, December 7, 2017

#AmAnth17 Wrap-Up: Anthropology Matters?



On Monday, I downloaded #AmAnth17 tweets.  This proved in many ways elusive and piecemeal.  First, the conference hashtags continue to shift.  Last year, the AAA finally discovered that the #AAA hashtag had other meanings and other audiences, among them AIDS activism in Japan and a pop music awards program in Korea (both of which prompted lively Twitter conversations this year).  Their efforts to promote alternative hashtags resulted in confusion, with people tweeting at #AmAnth17 (the ‘official’ hashtag), along with #AmAnth2017 (which would have been logically consistent with previous years) and, for the hell of it, #AAA2017. So the graph below includes tweets with any one of the three, with the top 50 Twitter users (by in-degree centrality) labeled. 



Here are the general metrics on this network.  

Graph Metric
Value


Graph Type
Directed


Vertices
426


Unique Edges
649
Edges With Duplicates
0
Total Edges
649


Self-Loops
145


Reciprocated Vertex Pair Ratio
0.047817048
Reciprocated Edge Ratio
0.091269841


Connected Components
100
Single-Vertex Connected Components
63
Maximum Vertices in a Connected Component
245
Maximum Edges in a Connected Component
445


Maximum Geodesic Distance (Diameter)
12
Average Geodesic Distance
4.810299


Graph Density
0.002783761



























It’s not an enormous graph, nor particular connected.  In many ways, it's similar to other graphs I’ve run in 2015 and 2016 (see.  For example, we see the same, prominent Twitter users.  Here are the top 50 accounts by in-degree centrality:

americananthro
profsassy
womenarchys
anthrofuentes
aprilmbeisaw
culanth
julielesnik
aba_aaa
sonyaatalay
hilaryagro
valorieaquino
altmetric
protest_matters
drtomori
jennyshaw011
lesleybartlett_
aunpalmquist
archyfantasies
twitatreyee
dukepress
stemethnographr
lauraellenheath
aaas_doser
pottershousedc
diane_tober
susangsheridan
blackfeminisms
afburialgrndnps
geekanthro
anthrosciences
stelynews
dcanthro
archpodnet
yarimarbonilla
oceaniajournal
beccapeixotto
mcclaurintweets
cvans
soclinganth
police_worlds
illinoispress
machristofides
anthroboycott
hildallorens
ratnagiri77
anthromuxer
tikabakic
camee_maddox

This is a great bunch of anthropologists and institutions, but, compared to previous years, Twitter traffic has diminished and, with it, topics have proliferated along lines of subdiscipline and sub-specialty.  That is, anthropologists (at least in their Twitter traffic) have retreated to the specifics of their panels and papers.  Here’s a word-cloud of the most frequently occurring 500 words from the 2017 AAA:


Now here’s another wordcloud from the 2015 meeting.  


The prominence of activist causes in 2015 (#BlackLivesMatter, BDS) stimulated tweets across subdisciplines in a way that is conspicuously absent from this year’s conference with some notable (and welcome) exceptions (thanks, @yarimarbonilla, @aba_aaa and others!).

Of course, these causes are still with us, along with a dumpster fire of authoritarian politics, fascism, rampant misogyny, ascendant white supremacy and environmental apocalypse.  But, in all of this, where does anthropology matter?  And if we can’t represent our united opposition to, say, fascist policies in the U.S., then what hope do we have of demonstrating the relevance of anthropology to anyone outside of this conference? 


In her critical summary of this year's meeting, Emma Louise Backe notes in Geek Anthropologist:  
the exclamatory nature of Anthropology Matters feels ineffectual. Are we trying to signal to the broader intellectual community and American public that anthropology does indeed matter? Or are we instead convincing ourselves that our choice of discipline was legitimate, necessary?
Well--those are the questions.  What will we--as anthropologists--do in the face of the palpable evil around us?  

Cybernetics and Anthropology - Past and Present

 I continue to wrestle with the legacy of cybernetics in anthropology - and a future premised on an anthropological bases for the digital.  ...