The diffusion of AI into everyday life via consumer products like smartphones and home assistants like Amazon Echo and Google Home has progressed to the point that they are now taken-for-granted presences, utilized readily and casually. For instance: my roommate kindly offers to walk my dog this afternoon and asks me what time would be best. I tell him between 1 and 2pm. Right away he addresses his Apple Watch: “Hey Siri — set an alarm for 1:30.” Why does he prefer this to setting an alarm manually? Giving the verbal command takes five seconds; typing in the command might take only 15 or 20, but that’s still three to five times as long. Also, he has a guitar in his lap. Giving the verbal command is physically easier. This is one small, almost insignificant, example of AI saving time by enabling multi-tasking. Multiply this over the course of a day, or a lifetime, and there’s the potential to save a lot of time. There’s also a feeling of empowerment, of being in charge, the conductor of one’s life, waving a verbal baton and seeing one’s desires fulfilled frictionlessly–or maybe occasionally having to enunciate more clearly. It helps the user feel more autonomous. Put simply: AI makes it easier to do things! To hear the news, to get the weather forecast, to turn on the lights, to hear a favorite song, and on and on.
The catch is that these technologies are bound up with what Harvard’s Shoshana Zuboff calls “surveillance capitalism.” In exchange for the utility provided by AI, users allow themselves to be surveilled, conceding much of their privacy and allowing themselves to be subject to the personally-targeted advertising that has supercharged the economic growth of Google and Facebook. But while that makes AI sound potentially dangerous for the user, the articles in this module indicate that most people accept this trade-off happily, and many people even feel positively about the advertising they see (Perdiman, 2018).
What does all this mean for libraries? One of the pillars of library values, articulated in ALA’s Library Bill of Rights, is the protection of privacy (ALA, 2019). Libraries provide a wealth of information to users while asking little of users in exchange. But does this also mean that libraries are less able to provide the personalized, data-driven service that people increasingly expect? For instance, by default most libraries do not keep records of users’ prior checkouts. Sometimes users come up to us and ask us for information on their past checkouts even though we don’t have it! We tell them they can enable tracking of their history in their Bibliocommons account, but we can’t see this history on our side. How do libraries compete with corporate information providers that make their business by catering to users’ personal desires and putting fulfillment right at their fingertips?
I liked Griffey’s distinction between the remote data analysis approach of Google and the local, device-based analysis used by Apple (2019). He is right to recommend that the local analysis approach should be used by libraries wherever possible. I’m not a budding systems librarian, but it seems to me a sensible suggestion by Griffey that libraries should take advantage of machine learning to make their offerings more “discoverable.” I think library collections can be analyzed and better promoted with the help of AI analysis without sharing private information remotely. It’s all a matter of library leaders making the correct calls when they’re negotiating with providers of AI technology. Hopefully, libraries can band together in new or existing consortia so that they can leverage AI in a cost effective way without resorting to contracting with large, for-profit companies.
What seems potentially tricker to me is how libraries can interface their content with consumer appliances like Echo and Google Home. Stephens (2018) suggests some intriguing possibilities, like syncing library audio content to these devices. While I was initially skeptical of this because such compatibility might seem to Amazon to pose a threat to Audible sales and subscriptions, there is a precedent for this kind of thing, since many libraries offer ebook checkouts that work with Kindle.
Overall, I think there are ways for libraries to integrate their services and content with consumer AI systems, but this is a path that must be tread carefully, making sure to avoid sacrificing the library’s important value of privacy.
ALA. (2019). Library Bill of Rights. Retrieved from: http://www.ala.org/advocacy/intfreedom/librarybill
Griffey, J. (2019, March 1). AI and machine learning: The challenges of artificial. American Libraries. Retrieved from: https://americanlibrariesmagazine.org/2019/03/01/ai-machine-learning-libraries/
Perdiman, D. (2018, January 5). Here’s how people say Google Home and Alexa impact their lives: People are increasingly willing to let go of their privacy concerns in exchange for a whole lot of convenience. Fast Company. Retrieved from: https://www.fastcompany.com/40513721/heres-how-people-say-google-home-and-alexa-impact-their-lives
Stephens, M. (2018, February 22). Flash briefing. Retrieved from: https://www.libraryjournal.com/?detailStory=flash-briefing-office-hours
Zuboff, S. (2019). The age of surveillance capitalism: The fight for a human future at the new frontier of power. PublicAffairs.