Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

AI, Insight, & Info Pros


Published on

Just as early online services didn’t just improve on the Reader’s Guide to Periodical Literature but revolutionized how information is accessed, so librarians and info pros need to look at AI from a different point of view and watch for opportunities to lead the AI discussion within their organizations. See an outline of my thoughts in this slidedeck.

Published in: Business
  • Be the first to comment

  • Be the first to like this

AI, Insight, & Info Pros

  1. 1. AI, Insight & Info Pros: Insight as a Service Mary Ellen Bates Oct. 22, 2019
  2. 2. Tweeting this? #InternetLibrarian @mebs Slides at: 2
  3. 3. When I knew we were doomed Your flight is scheduled to depart on time but due to a delayed incoming flight, there is a good chance it will be delayed by at least 30 minutes. Your flight is scheduled to depart on time but due to a delayed incoming flight, there is a good chance it will be delayed by at least 30 minutes. 3
  4. 4. AI: What do you think of? Smart speakers Amazon / Netflix recommendations Chatbots Facebook recognizes friends’ faces in photos Self-driving cars IBM Watson 3CPO / Her 4
  5. 5. AI is a toolbox Machine translation Text analysis Robotics Deep learning… 5 Computer vision Facial recognition Speech recognition Pattern recognition
  6. 6. AI isn’t thinking (yet…) AI can find unexpected patterns Humans can: Create new product Design something a new way Question why we do something Respond to unrecognized needs 6
  7. 7. What AI can’t do AI can’t empathize [yet] AI can't get creative AI can’t tell you the so-what? AI can’t suss out the “question behind the question” Info pros can stay two steps ahead 7
  8. 8. AI is ALREADY in the library Enhancing discoverability: Entity recognition Image and audio recognition Taxonomy development Sentiment analysis Recommendation engine 8
  9. 9. “How can AI help do my work better?” is the wrong question Transformative tech doesn’t do what we do better It does something we can’t do 9
  10. 10. Librarians Can OWN the AI Discussion 10
  11. 11. Serve as AI interpreter and guide Univ of Rhode Island Library’s AI Lab ( AI workstations (with open source tools) Curate open data sets Host meet-ups, workshops and discussions Curate online courses, journals, other resources 11
  12. 12. Library AI concerns Reliance on deep learning Do we trust algorithms we can’t understand? Need to teach info literacy Why do we trust Google more than Wikipedia? Deep learning vs transparency 12
  13. 13. Library AI concerns Need to advocate for “explainable AI” systems Demand transparent processes How do we negotiate content licenses, partner with info providers? 13
  14. 14. Library AI concerns How do we evaluate and acquire OA content? How can we leverage and share our special collections? Do we have the resources to digitize? Should we? 14
  15. 15. Impact of AI on libraries Do libraries have to hire data specialists? Do we provide analysis or “just” the tools? How do libraries keep staff AI-literate? 15
  16. 16. Impact of AI on libraries Will this change the library’s form and function? Design for people to view the info or for AI to access the info on behalf of humans? Do users bring their own tools and access the digital version of the library? 16
  17. 17. No answers, just lots of questions! 17
  18. 18. Slides at Mary Ellen Bates @mebs +1 303 772 7095 maryellenbates 18