This is a thrown-together post to capture responses to a question I asked on twitter last week. The Digital Scholarship Reading Group I run at the British Library will spend the first meeting of 2021 collaboratively planning topics to discuss in the rest of the year, so to broaden my understanding of what might be discussed, I posted, 'A question for people interested / working in Digital Humanities – what do you think are the big topics for 2021? Or what's not, but should be a focus? … New publications or conference papers welcome!'.
And since I was asking people for suggestions, it seemed like the right time to share something we'd been thinking about for a while: 'we've decided to open our discussions to people outside the British Library / Turing Institute! We'll alternate between 11am-12pm and 3-4pm meeting times on the first Tuesday of each month'. I haven't sorted the logistics for signing up – should it be on a session by session basis, or should we just add people's email address to the generic meeting request so they get the updates? (Will they get the updates, given how defensive and awful email is for collaboration these days?)
I also posted links: 'For context, here's what we read up to early 2018 What do deep learning, community archives, Livy and the politics of artefacts have in common? and a themed summary, Readings at the intersection of digital scholarship and anti-racism.
Responses to date are below. I didn't want to faff about with embedded tweets because they're more likely to break over time, so I've just indented replies with the username at the start.
Claire Boardman @boardman_claire The environmental impact of DH? Conversational AI and collections?
Jajwalya Karajgikar @JajRK Large language models, and computational text analysis overall?
@mia_out As in models that use very large amounts of training data? And yes, we should do more on CTA, I think we could probably get broader coverage of methods, thanks for the prompt!
Jajwalya Karajgikar @JajRK Models that use deep learning for language prediction; GPT-3 I think someone mentioned on the thread already?
Thomas Padilla @thomasgpadilla Social justice and DH – though all work that frames current strife as a new thing vs. a longstanding pervasive reality should be tossed into an abyss to make way for others
@mia_out I won't ask you to name and shame bad pieces, but let me know if you have any favs that do it well!
Thomas Padilla @thomasgpadilla Ha! On the collections side @dorothyjberry has a piece or two brewing. @ess_ell_zee work here is good too I think https://journal.code4lib.org/articles/14667
@artepublico peeps like @gbaezaventura and @rayenchil and the @MellonFdn supported Latinx DH program are good places to look
Same goes for @profgabrielle and all the @CCP_org work is fantastic
Wilhelmina Randtke @randtke Long term sustained funding. Acknowledging, and even compiling a list of, projects that have had resources eliminated or been completely discontinued since March.
Jenny Fewster @Fewster Absolutely! This is a problem internationally. Dig hums projects set up with one off funding that then aren’t sustained. Unfortunately digital projects are not a “set and forget” prospect. It’s a colossal waste of time, effort, knowledge and money
Matthew Hannah @TinkeringHuman I think we need/will see more work about the limits of neoliberal capitalism, the academy, and DH, applications of critical university studies and Marxist theory. Esp as higher ed continues to implode.
@mia_out Sounds very timely! I don't suppose you have any papers or presentations in mind?
Matthew Hannah @TinkeringHuman Claire Potter’s piece in Radical Teacher is also an inspiration: https://pdfs.semanticscholar.org/c3c0/b0f853710a56b13b0d232b3b435a19bf59a7.pdf
But we need more engagement I think around the question of precarity and economics imo
Johan Oomen @johanoomen Detecting polyvocality in heritage collections and navigating this underexplored dimension to investigate shifting viewpoints over time. Could also be a great opportinity for crowdsourcing projects, to encourage contemporary users to voice their opinions on contentious topics.
@mia_out Ooh, that's a really juicy one – lots of potential and lots of pitfalls
Erik Champion @nzerik The influence of social media on politics? The failure of social media apps, webchat etc to compensate for lockdown distancing? Govt and corp control on personal data? Big companies controlling VR devices and personal +physiological data?
@mia_out As seen recently when people were annoyed they had to do a Google Recaptcha on a COVID test site
Erik Champion @nzerik Bots need vaccines too! (Equality for bots trojans and spam machines #101)
Alexander Doria @Dorialexander On the technical side, optical manuscript recognition and layout analysis (especially for newspaper archives): mature tools are just emerging and that can change a lot in terms of corpus availability, research directions and digitization choices.
@mia_out There is so much interesting work on newspapers right now! It feels like scholarship is going to have a quite different starting point in just a few years. Periodicals less so, maybe because they're more specialised and less (family history) name rich?
Alexander Doria @Dorialexander Yes that's true. Perhaps also because they are less challenging both technically and intellectually (it's not that much of a stretch to go from book studies to the periodicals).
Alexander Doria @Dorialexander (On the social side I would say there is a long overdue uncomfortable discussion about the reliance of the field to diverse forms of digital labors: from the production of digitized archives in developing countries to the large use of students as a cheap/unpaid labor force)
@mia_out That ties in with ideas from @TinkeringHuman
Max Kemman @MaxKemman I think we'll be seeing more about Computational Humanities and how it relates to Digital Humanities, for which a good starting point will be the @CompHumResearch conference proceedings http://ceur-ws.org/Vol-2723/
@mia_out Ooh, we could have a debate or discussion about the difference!
Lauren Tilton @nolauren And intersection/ difference from Data Science
@mia_out Good call, the lines are becoming increasingly blurred, hopefully in more good ways than bad
Gabriel Hankins @GabrielHankins GPT-3 and algorithmic composition. Interested in the conversation if you open it!
And finally, one reason I collected these responses was:
Michael Lascarides @mlascarides A feature I wish Twitter had: When I see someone influential in a domain I'm interested in ask a really great question, I want to bookmark that question to return to in a couple of days once the responses have come in. It's a use case a bit more specific than a "like".
Michael Lascarides @mlascarides Inspired most recently by [my] Q, but it comes up about once a week for me
Additional responses, firstly from Alex Hailey: 'Coming from a library perspective, I'm going to borrow the phrase 'the imperative to broaden one's dataset' from a recent Yael Rice & Sonja Drimmer article and say that I would like to see more focus on the decisions and factors which shape digitised collections, and training datasets used in ML, and the potential for more collaborative work between GLAMs and researchers to address this'
And a summary of the group's discussion today: 'cultural analytics, decolonising archives, horizon scanning for emergent technologies that will affect libraries and library users, user experience (including mental models that people bring to different services / methods / data), the effect of social media on politics and news, access to subscription data (ProQuest, Gale etc) and open datasets to apply digital methods, practical ethics / evaluating AI systems, overviews of humanities and social science approaches for IT folk'.