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.

Robots: What Could Go Wrong? What Could Go Right?


Published on

A presentation given at the ALA Midwinter Conference, Philadelphia, PA. Jan. 26, 2020 by Bohyun Kim, CTO/Associate Professor at the University of Rhode Island Libraries.

Published in: Technology
  • Be the first to comment

  • Be the first to like this

Robots: What Could Go Wrong? What Could Go Right?

  2. 2. THIS IS A “DISCUSSION" SESSION. ▸ Discussions are designed to spark conversation across participants. ▸ Lead discussants or facilitators poses questions or prompts and encourage attendees to share their perspectives and insights. ▸ These sessions might be especially useful for early-stage exploration or community-building around new and emerging ideas.
  3. 3. WE WILL DISCUSS… I. What counts as a ‘Robot’? II. General concerns about the adoption of robots III. Robots for libraries * Need a note-taker, a time-keeper, & a question-tracker * Google doc at
  4. 4. I. WHAT COUNTS AS A ROBOT? (ice-breaker)
  5. 5. 1. ‘ROBOT’ ▸ Appearance like a human? ▸ A physical body? ▸ Intelligent / smart? ▸ No agreed-upon definition
  6. 6. CANDIDATES ▸ ‘robota’ - Originally meant “forced labor” in Czech. ▸ Anything that can be programmed to move, no matter how simple? ▸ Software designed to automate tasks? ▸ Usually people want to distinguish a robot from a piece of software or a typical machine.
  7. 7. A WORKING DEFINITION ▸ “A machine, situated in the world, that senses, thinks, and acts” — from George Bekey, “Current Trends in Robotics: Technology and Ethics,” in Robot Ethics: The Ethical and Social Implications of Robotics, eds. Patrick Lin, Keith Abney, and George Bekey (Cambridge, MA: MIT Press, 2012), p.18.
  8. 8. ‘EMBODIED’ ▸ Means that it should be able to take in sensory input from the surroundings; processing it for a cognitive task; and acting upon the physical world. ▸ Need sensors; ▸ Need some cognitive capability to process the input from the environment and to determine the response or an action; ▸ Need actuators, which allow it to physically act upon the environment.
  9. 9. AI-POWERED ROBOTS ▸ AI has made it possible to make a robot intelligent and autonomous in performing tasks that are not only mechanical but also cognitive. ▸ e.g. driving, natural language processing, translation, and face recognition. ▸ The capability of AI-powered robots far exceeds that of other simpler and less sophisticated machines.
  10. 10. ROBOTS & HUMANS ▸ Robots have no consciousness, desire, motive, or other mental states. ▸ Humans have the tendency to anthropomorphize many things, including robots. ▸ The adoption of robots raises interesting and unique concerns.
  12. 12. 2. GENERAL CONCERNS ▸ Safety ▸ Morality (*Trolley case for an autonomous vehicle) ▸ Impact on social relationships ▸ Potential to be used as a means for manipulation and deception
  13. 13. 1. What kind of safety mechanism(s) should a (library) robot be equipped with? 2. How should an AI-powered trolley act? 3. What kind of relationship is likely to form between people & (library) robots? 4.How may robots be used to manipulate people & how can we design robots to prevent such potential manipulation? [15min] 1:10-1:25 PM
  14. 14. DISCUSSION (A) SAFETY ▸ How to prevent robots from harming humans either by accident or while trying to achieve its goal by design? ▸ A kill switch & an override option
  15. 15. THE CONFIGURATION OF A ROBOT FOR SAFETY ▸ Depends on the specific type of task that it is supposed to perform & the risks associated with the given task.
  16. 16. LEVELS OF AUTONOMY ▸ ‘human-in-the-loop’ ▸ ‘human-on-the-loop’ ▸ ‘human-off-the-loop’ ▸ The efficiency and the cost-savings and the safety concerns can come into a conflict and need to be balanced.
  17. 17. DISCUSSION - (B) MORALITY ▸ Morality becomes a questions with the AI’s decision-making capability. ▸ To what use should we put those AI systems (/robots) & how much control should we allow them to have? ▸ An important question as seen in the trolley problem.
  18. 18. TO ADDRESS THE ISSUE OF MACHINE MORALITY, ▸ Determine what level of autonomy and ethical sensitivity a robot is to be equipped. ▸ Consider what level of machine morality may be feasible and appropriate for the given robot. ▸ Considering their embodied nature and high level of autonomy, we should start investigating how our relationship may and should look like with these robots as the members of our society.
  19. 19. DISCUSSION - (C) HUMAN-ROBOT RELATIONSHIP ▸ Humans have the tendency to project human qualities onto a robot and form an emotional attachment. ▸ Can be beneficial if those help to fulfill the robot’s intended use. (e.g. trauma patient) ▸ Can be harmful to the robot user and impede its intended function. (e.g. a military ordnance disposal robot)
  20. 20. INTENDED USE ▸ Those who build a robot should consider ▸ what type of relationship is the robot supposed to form with its user for optimal functioning? ▸ Robot users need to be aware of the fact ▸ that a robot may be designed to elicit anthropomorphic projection to perform its function. ▸ Is a robot to be regarded as a social agent?
  21. 21. DISCUSSION - (D) MANIPULATION ▸ More robots will be introduced in a personal environment such as homes and care facilities, playing a social role as a companion or a caretaker ▸ Many will treat them as social agents either consciously or unconsciously. ▸ But robots are designed by robot manufacturers, who seek to maximize their profit, and may become a tool for manipulation.
  22. 22. ABUSE/DECEPTION BY ROBOT MANUFACTURERS ▸ Charging an exorbitant amount of fee for a care robot’s software upgrade ▸ Programming a robot to suggest a purchase to its user ▸ Mishandling personal and private information that a user confides to the robot
  23. 23. ISSUES ▸ Manipulation and deception through a robot is particularly pernicious because it preys on people’s natural inclination of caring about other social agents. ▸ Raises the question of whether a robot is to be treated as a social agent or a mere tool/thing?
  25. 25. 3. ROBOTS FOR LIBRARIES ▸ AuRoSS, a robotic shelf scanning system, Singapore ▸ Nao robots - Westport Public Library, USA ▸ Finch robots - Chicago Public Library, USA ▸ Robot Day - San Diego Public Library, USA
  26. 26. 1. What kind of robots we may come to see at libraries in the near future? 2. What kind of human-robot interactions may take place at libraries? 3. What type of human-robot relationship may facilitate or impede a library robot’s intended function? [15min] 1:35-1:45 PM
  27. 27. AREAS FOR ROBOT ADOPTION ▸ Greeting & Directions ▸ University of Pretoria, South Africa - ‘Libby’ ▸ Access Services (e.g. AuRoSS) ▸ Reference ▸ Alexa at the University of Oklahoma, ▸ Reader's advisory service ▸ A reading robot - ‘Luka’
  28. 28. HUMAN-ROBOT INTERACTIONS AT LIBRARIES ▸ Library robots play the role of an assistant and a companion that help library users with accomplishing a variety of tasks. ▸ The resulting relationship is likely to be positive and friendly. ▸ Library robots as mere tools or more? ▸ Should libraries still encourage library patrons to form a bond with a library robot? ▸ How would the previously-discussed concerns apply to library robots?
  29. 29. A LIBRARY IS/DOES… ▸ A safe space mentally and physically ▸ Free/equal access to information to the public ▸ Empower people through knowledge. ▸ Protect people’s intellectual freedom. ▸ Help people exercise their right to pursue information and knowledge privately without being monitored or surveilled by a third party.
  30. 30. GUIDELINES FOR LIBRARY ROBOTS ▸ Would library robots be able to fulfill the library’s mission just like librarians? ▸ Library robots, particularly those that will be involved in library patrons’ information-seeking activities, may have to follow stricter guidelines than other robots deployed in non- library areas.
  31. 31. IN THE FUTURE, ▸ More sophisticated, versatile, and autonomous robots are likely to enter our homes, workplaces, and libraries. ▸ Neither our society nor the library yet fully understands how the wide adoption of robots will affect us. ▸ As a new type of social agent, robots will generate a lot of interesting questions.
  32. 32. FOR MORE DETAILS, SEE ▸ Bohyun Kim, ”AI-Powered Robots for Libraries: Exploratory Questions." IT Satellite Meeting, IFLA WLIC Conference, 21-22 August, 2019, Wildau, Germany. Conference Paper. ▸ Full text available at: lib_ts_pubs/113/ (URI Digital Commons repository) or
  33. 33. Questions?
  34. 34. THANK YOU! @bohyunkim