TILT #56 — Your algo sucks, your people aren’t always better

Jessamyn West
today in librarian tabs
5 min readMay 10, 2018
image from Central New York Library Resources Council Archives

I feel like this newsletter should open with “What crazy algo shit is happening this week?” Google launched their assistant thing that makes phone calls, while making “ums” and “hmms” pretending to be a human. So many questions!

First and foremost is “Is it fair to use human-sounding AI robots to call actual humans without disclosing that’s what’s happening?” Good question, says Google, we’ll handle that later.

The Google Duplex technology is built to sound natural… It’s important to us that users and businesses have a good experience… and transparency is a key part of that. We want to be clear about the intent of the call so businesses understand the context. We’ll be experimenting with the right approach over the coming months.

Librarians, now is your chance for some brand disambiguation: Libraries, we’ll never make you talk to a robot!

Unrelatedly, a friend searching Google for organized crime knock knock jokes (for reasons) found some unexpectedly racist results. Search engine optimization is a shady business. But hey at least MartinLutherKing.org is off the front page, amirite?

My class wrapped up at the University of Hawai’i and I was looking for some nice beach photography to put in an email. I went to DPLA and searched for hawai’i beach. I found two interesting things.

1. Some of the images from NARA were loading into the thumbnail pane as full-sized images. I’ve spoken before about how the internet at my house was broadband but isn’t anymore. Seeing a page of half-loaded images gave me flashbacks to the dial-up days.

I messaged some folks at DPLA who said this was because of the data NARA sent them and then they fiddled with stuff a little. I just got word this week that they fixed it! So nice.

2. I found an image I wanted to use in the University of Southern California’s special collections. I can be a broken record about content archives being rights-searchable to facilitate better sharing (Google Images has this, Trove has this, Europeana has this)

This image had an unclear-to-me rights statement. It was public domain, but also released under a CC-BY license? Links in the metadata just went to catalog searches for items with those keywords (a buggy feature I do not love). Look at this cute Department of Public Works sewer van that I found when I clicked on the word “public” hoping to learn about licensing.

image from USC Libraries and California Historical Society

I emailed the library who confirmed this was, indeed, what that statement meant, stating “the original image is in the public domain, but you are not looking at the original image.” Well.

This is sort of a thing libraries do, take a public domain image and put a more restrictive license on the digital surrogate they create. Peter Hirtle wrote up an analysis of this practice over a decade ago which still sounds modern. I can understand the impulse to not want to allow wholesale scraping of free images. But then we’re back to algos again, worrying that corporate “for sale” versions of public domain images will be easier to find than ours, and thus not really increasing shareability after all. It’s a conundrum.

A few personal notes this week. I performed a wedding at Boston Public Library this past weekend (for some book people) which was basically a dream come true. It was interesting seeing library-as-costume as opposed to library-as-workplace or library-as-hangout. This antique card catalog held people’s dinner seating assignments.

On the way back home, I did my privacy talk at the tail end of Choose Privacy Week at the Brooks Library in Brattleboro. I also contributed a blog post to the CPW blog talking about how and why I began to give that talk.

If you’re within a few hours of me and want someone to come talk with your patrons about privacy, I have incredibly reasonable rates — $50 + mileage + dinner — look me up.

Some odds and ends:

Errata: I somehow managed to mess up the timeline of Hawai’i in the last newsletter. It’s been corrected in the online version.

Today in Librarian Tabs is written irregularly by Jessamyn West who also maintains librarian.net. It’s available in more-accessible format your inbox via TinyLetter. Thanks for reading.

--

--

Jessamyn West
today in librarian tabs

Rural tech geek. Librarian resistance member. Collector of mosses. Enjoyer of postcards. ✉️ box 345 05060 ✉️ jessamyn.com & librarian.net