A newsletter to learn about practical ways to decolonize your research and data work-lives with byte-sized drabbles about the daily life of a data professional.

Somehow I did A.I. projects the last 2 years…

This week wraps up my data science fellowship project from last year. I presented my Relational Possibilities Project one final time at the Digital Library Federation Forum with collaborator, kYmberly Keeton. Bro, I am tired. Look at all the stuff we’ve done with this project:

Relational Possibilities was fun. I felt I could experiment with different data science tools and career paths and I ended up doing more artificial intelligence research, got into streaming on Twitch, and made several video games and applications because of this project.

I started a long-term A.I. project to show emotional labor of archivists through A.I. imagery

I started down the A.I. path, a couple years ago when I answered a call for a public history conference in Malta. I ended up giving a speech on the history of emotional labor in the United States archival professional with my long-time collaborator, Kristen J. Nyitray.

At the time I felt frustrated and disillusioned with data work and wanted to explore these emotions as societal expectations and outcomes. In 2022 I published a research study in this corporate/private sector journal for digital asset management (DAM) where I analyzed job ad data for DAM work and found there was literally a $20,000 pay discrepancy if you got a government/public job versus worked for a corporation or foundation. That pissed me off!

That’s more to it - like corporate jobs typically asked for less qualification and education and had more flexible work arrangements - but the fact that public sector jobs for data required you to hustle more for significantly less money was infuriating! There is no official DAM degree or career path, so the equity issues are a direct result of employer’s subjective definitions of value for this kind of data work.

This prompted Kristen and I, who share similar experiences in archival work, to research why this is the way it is today. We found what we expected: feminized labor in data is devalued.

Kristen and I recently presented at another public history conference in Northern Ireland and showed A.I. images we generated using dadaism in the prompt. The point was to get at our own emotions when we think about these issues of emotional labor while simultaneously seeing what A.I. data for this specific model knows about our emotions.

I’m really interested in A.I. through the dadaism lens because as an art form dadaism expresses nonsense and anti-bourgeois protest. Similarly, AI critiques our data-driven world by exposing biases and glitches within the eerie beauty of machine-generated chaos. To me the imagery is also very horror genre coded. I love using horror as a medium to explore deep societal issues.

I plan to continue this archives and emotions research into 2025, but also try less A.I. things. A.I. isn’t that useful for the kind of data work I do and I question the utility of A.I. for most data professionals. From what I have seen, most A.I. sells enhanced productivity for problems that aren’t real. Case in point: a A.I. transcription service for a Zoom recording doesn’t seem useful because you can watch a video recording. It doesn’t replace experiencing the meeting/event.

Let’s Create!

I’m open to collaborations, freelance gigs, and conversations about the ideas I shared. Feel free to get in touch and comment on the newsletter.