In southwestern Japan, there is a hotel staffed primarily by robots. Royal Caribbean’s new Anthem of the Seas cruise ship has a bar staffed by robot bartenders. And Starwood’s Aloft hotel in Cupertino, CA, uses robot butlers. With this type of technology beginning to appear in the hospitality industry, libraries might begin considering the implications of patron interactions with service robots in a library environment, Jamie Hollier, coowner and co-CEO of technology consultancy Anneal, noted during the Library Information Technology Association’s (LITA) Top Tech Trends panel at the American Library Association’s (ALA) 2016 Midwinter Meeting in Boston.
The robots were mentioned in a brief aside as Hollier made a broader point about autonomous agents, computer programs that conduct operations for a user based on environmental cues, previously defined user preferences, and other information. She cited “Amy,” the personal assistant/scheduling bot by New York–based start-up x.ai, as an example.
“You can use natural language and cc [Amy] on emails when you’re trying to set up meetings, or you can email her directly and say, ‘Set up a meeting with this person on this date,’” Hollier said. “She basically functions as your assistant and emails them and sets up the schedule…. She has access to your Google calendar…. Here’s the thing about her: I’ve been using her a little bit, and I decided to stop, because she’s kind of creepy!” Hollier said. The capacity crowd laughed, but she went on to explain that one of her clients had written her to compliment Amy, assuming that the program was human.
“She’s not [artificial intelligence], she’s a logic engine, but she uses natural language really well,” Hollier said. “These things are going to continue to be more prevalent in all different places, so what happens when libraries start using autonomous agents for their emails to [patrons] about overdue books, or something like that?”
Hollier was joined by moderator and 2012 LJ Mover & Shaker Lisa Bunker, social media librarian for Pima County Public Library, AZ; Jason Griffey, founder and principal of consulting and creation firm Evenly Distributed and a 2009 LJ Mover & Shaker; Jim Hahn, orientation services and environments librarian and associate professor at the University of Illinois Undergraduate Library; Alex Lent, director of the Millis Public Library, MA; Thomas Padilla, digital scholarship librarian at Michigan State University Libraries; and Ken Varnum, senior program manager for discovery, delivery, and learning analytics at the University of Michigan Library.
“Autonomous agents are becoming frighteningly good,” Griffey agreed. “The meta trend on top of that is generalized machine learning being ridiculously good…. There’s been research done that shows that machine learning systems, when fed patient information, are more accurate in diagnosing illness than doctors are. Obviously, we all know about robot cars—cars are a system where humans are just about the worst possible answer for what should be driving [them]…. This is a machine learning algorithm that’s getting better and better and better and better.”
Open source for savings
In a presentation that later drew several queries during the Q&A portion of the session, Lent discussed his small library’s decision to switch to open source software, including the Ubuntu Linux operating system.
“Open source technology is generally free, and when your budget for technology doesn’t exist, that’s really a big perk,” Lent said. The library also switched from Microsoft Office to open source LibreOffice suite and launched a new website using the open source WordPress content management system. Lent said that upkeep, troubleshooting, and maintenance tasks often fall on him but noted that open source programs and systems tend to have robust support communities online, and it is often easy to get quick responses by posting questions or problems on Twitter. When audience members asked whether patrons might be upset or disappointed about the lack of access to industry-standard software such as Microsoft Office, Lent said that it has been six months since the library made the switch, and he has yet to receive any complaints.
In response to a separate question, Lent said the library had not considered migrating to an open source ILS. Millis is part of the Minutemen Library Network, whose members use Innovative Interfaces’ Sierra Services Platform.
Location, location, location
Hahn discussed the growth of location-based information services, such as indoor positioning systems to help patrons navigate the stacks, or Bluetooth beacons, which several libraries have begun using to send targeted, opt-in messages to the smartphones of patrons who visit specific areas of the library.
“The book stacks offer a really fascinating area for location-based services…largely due to intellectual organization by shelf classification, and the already-existing collocation” of related books, Hahn said.
Analytics and privacy
Gathering data on how college students use campus and library resources can help libraries quantify their contribution to campus and help an institution address barriers to learning and improve student outcomes. However, gathering data, especially data that could be tied to individual students, raises a host of privacy concerns, noted Varnum. Many university libraries have long-standing policies that go as far as regularly purging checkout histories and clearing public computer event logs in an effort to protect privacy.
The University of Michigan has been working on a campuswide learning analytics program that protects student confidentiality and attempts to strike a balance between privacy and the collection of data that could help the university enhance the experience of its students.
The term praxis—which is defined as the practical application of a theory—has become popular among data scientists as a way to describe how insights gleaned from data can translate into actionable ideas. However, Padilla noted that the term has long been applied within individual disciplines, such as chemistry or the social sciences. He is anticipating the emergence of what he described as “broad data praxis,” which is not confined to any specific disciplinary community.
“One real-world example of broad data praxis is data-driven journalism—interactive, web-based visualizations accompanying in-depth reporting on police violence in the United States,” for example, he said. “Because journalism is generally geared toward communicating information broadly to a diffuse community, it holds the potential to naturalize and normalize aspects of working with data…. [such as] data provenance, data quality, data representation, data documentation.” This trend may have interesting connotations for how libraries and universities approach data literacy, Padilla said.
Block chain potential
In addition to his comments regarding autonomous agents and machine learning, Griffey also discussed block chain technology. A block chain is a permissionless, distributed database system that was originally developed to validate and record transactions in the bitcoin cryptocurrency network. When users transfer bitcoin currency to other users, a time-stamped transaction is publicly broadcast to the bitcoin network. Ledgers of all transactions are then stored in cryptographically enforced “blocks” several times per hour, which are then published throughout the network, creating a sequential time-stamped record “chain” showing who owned what and when.
This methodology reconciles millions of anonymous transactions in a true peer-to-peer environment that does not rely on any central authority, Griffey noted. “The underlying technology is incredibly powerful…. And at its base the block chain is just a ledger system—it’s literally a system of who gave what to who—which sounds a whole lot like what library systems do.”
Interoperability on the rise
Hollier also discussed the growth of interoperability, with application programming interfaces (APIs) facilitating the transfer of data among programs or systems, along with efforts to standardize data formatting within specific industries. She outlined her thoughts on this topic at length in a recent blog post.
“It’s nothing new. It’s been happening for a while, but it’s starting to get traction in new and different ways,” she said. One example is FHIR (Fast Healthcare Interoperability Resources). Historically, medical providers and insurers have used a variety of proprietary codes and cataloging methods when recording information on patients, leading to fragmented records and problems when records of different providers need to be merged. FHIR is an effort to create data standards within the medical industry that was driven by new requirements for electronic medical records in the Affordable Care Act.
Separately, software containers, such as Docker, are beginning to facilitate interoperability by making it possible for apps and software tools to overcome fundamental differences in infrastructure and language.
“What that means is that people can start sharing applications and software in a whole new way,” Hollier said.