keynote for CILIP Health Libraries Group virtual conference - 24th November 2021

1. Health information professionals
and Artificial Intelligence
Andrew Cox, Senior Lecturer, Information School, Sheffield
a.m.cox@sheffield.ac.uk
The impact of AI, machine learning, automation and robotics on the information professions: A report for CILIP
https://www.cilip.org.uk/general/custom.asp?page=researchreport
Health Libraries Group – Virtual conference - 24th November 2021

2. The Information School
University of Sheffield, UK

3. Artificial Intelligence

4. Part 1: Defining AI

5. Definition 1: Everyday functionality increasing
productivity of knowledge workers
• Auto-correct
• Auto-suggest
• Grammar check
• Transcription/ captioning of
meetings
• Translation tools
• Automatic writing
• Many features of search and
recommendation
• Increased productivity
• Increased access to knowledge
• Does it do so evenly, eg which
languages are served best?
• Continued need for critical
information literacy

6. Definition 2: A cluster of new technologies
• Machine learning
• Natural language processing
• Sentiment analysis
• Voice recognition
• Image recognition
• Applications: new forms of search and
recommendation, collections as data
• Eg Searching medical images or text
mining the literature
• Chatbots, voice assistants
Its not magic
• Data description and fragmentation
• Training data effort
• Data bias
• Data validity
• Data ethics
• The hype around technological
solutionism always downplays
women’s work, emotional labour,
service (Mirza and Seale, 2017)
https://exploreai.jisc.ac.uk/

7. Definition 3: A global industrial complex
• How much energy does it use?
• How much raw material does it
use?
• How much unpaid labour/ data
are we contributing?
• How much does it rely on
clickworkers?
• Who profits? How does it link to
global inequalities?
• Is it sustainable?
(Crawford and Joler, 2018; Crawford, 2020)

8. Definition 4: An evolving idea… a social
imaginary
• Utopias and dystopias of science fiction
• Evolves as our idea of technology and intelligence evolve
• Intelligence as answering a quiz
• Intelligence as social, creative, emergent, open

9. Definition 5: AI for information professionals
Use in our own daily work
• Using AI tools in your own work, eg book ordering,
summarising texts… etc
• Robotic process automation of backend processes
Using AI in our own services to users
• Supporting use of AI built into third party search
tools
• Applying machine learning to organisational datasets
• Text mining the literature
• Developing a chatbot or voice assistant for
information service users
Supporting a community of data scientist users
• Locating/ licensing external data sets for analysis
• Developing training data for machine learning
• Creating a community of data scientists
• Data governance and stewardship
Supporting wider use in the organisation
• Input to your organization’s adoption of AI, eg in
procurement, training etc
• Promoting data and AI literacy
Using it to monitor/ predict/influence user behaviour
• Applying sentiment analysis to social media
responses to information services

10. What types of uses can you
imagine?
Post your answers in padlet: https://padlet.com/a_m_cox/jduwltdylkl7vi54

11. Understanding or managing users
• Libraries are rich in data about users
• Turnstile data
• Circulation data
• Usage of digital resources
• Satisfaction surveys
• Reference enquiries
• Qualitative data, eg UX studies
• AI to analyse social media data or open text survey data
• AI to nudge users?
Learning analytics debate
• Lack consent or student awareness of how their data
is being used
• Lack ethics review
• Libraries few responsible use statements
• Benefits unclear or for institution not learner
• Privacy issues
• Chilling effect on free speech and expression
Jones et al. (2020)
• Issue of validity: “learning data” not equal to learning
• Evidence that data analysis flawed (Robertshaw and
Asher, 2019)

12. Ethical AI

13. Ethics and values – do our profession’s
translate?
• Study found 84 statements about
ethics (international organisation,
governments, tech companies etc)
(Jobin et al., 2019)
• Bias
• Transparency and accountability
• Privacy / surveillance – chilling
effects
• Security and safety
• Human agency - nudging
• Voice assistants cannot understand
people with non-standard accents
• “Female” naming of chatbot
reinforces stereotypes
• Deepfakes create false news
• Automated moderation on social
media blocks legitimate free
expression
• Fear of surveillance inhibits
searching/ reading certain material
• My data being used without my
consent

14. Which of these barriers to using AI do you
think is the greatest?
1. Privacy concerns
2. Concerns about bias and lack of transparency
3. Security risks
4. GDPR compliance
5. Legal uncertainty
6. Issues of data quality in general
7. Over promising/ hype
8. Lack of information professional skills
9. Other people’s dated assumptions about what information professionals can do
10. IT owns the agenda
11. User fear of technology/ need for culture change
12. Other priorities

15. Part 2: AI and information
professional jobs

16. AI’s impact on jobs
• Replaced
• Dominated
• Divided
• Complemented
• Augmented
• Rehumanized
(GPAI, 2020)
Differences by sector
• Some scholars predict the
replacement of professionals
(Susskind and Susskind, 2015)…
others emphasise the
vulnerability of “routine” jobs
(Frey and Osborne, 2017)…

17. Skills and knowledge – updating for data and AI
• Information Management skills – apply to data – information governance,
standards, metadata, preservation
• Sourcing external data is a content challenge
• Collection management, including metadata, standards, IPR etc – data not print
collections
• Importance of data provenance for validity
• Procuring data and systems – different types of system
• Searching for data – new landscapes of search
• Teaching data & AI literacy – new dimensions of literacy, implies stronger
understanding of data analysis
• Knowledge of users’ need – assembling data from wide range of sources
• “Computational sense” (Twidale and Nichols, 2008)

18. Attitudes
• Service focus / balanced with sense of institutional agendas
• Collaborative skills / Influencing skills
• Commitment to professional development and learning
• Professional knowledge sharing

19. WEF (2020) – 10 skills to thrive in the 4th
Industrial revolution
1. Complex Problem Solving
2. Critical Thinking
3. Creativity
4. People Management
5. Coordinating with Others
6. Emotional Intelligence
7. Judgement and Decision Making
8. Service Orientation
9. Negotiation
10. Cognitive Flexibility

20. Addressing our weaknesses
• Other professions who talk more about data:
• I do not think we need to become data scientists (who combine stats/
computing/ subject knowledge)
• But other groups such as BCS, DAMA also talk about data management
• Literate profession – shift towards quanti data (according to
McKinseys)
• Lack of recognition for IM/KM
• When we see tech as outside force – IT control the agenda
• Not so good at collaborating across sectors
• Our diversity

21. Values
Example: CILIP’s ethical statement
1. Human rights, equalities and diversity, and the
equitable treatment of users and colleagues
2. The public benefit and the advancement of the
wider good of our profession to society
3. Preservation and continuity of access to knowledge
4. Intellectual freedom, including freedom from
censorship
5. Impartiality and the avoidance of inappropriate bias
6. The confidentiality of information provided by
clients or users and the right of all individuals to
privacy
7. The development of information skills and
information literacy
Offers a distinctive perspective on AI ethics
• Understanding of how inequality
reproduced?
• Relevance of sustainability, eg need
for green AI
• Consideration of global South
perspectives, eg issues around
language and operating in low
resource environments
• The ultimate drivers for
datafication lie beyond our control

22. A guide to good practice for digital and data-
driven health technologies (2021)
• Review the Data Ethics Framework and abide by the principles
• Ensure that the product is designed to achieve a clear outcome for users or the system *
• Ensure that the product is easy to use and accessible to all users *
• Ensure that the product is appropriately tested and is fit for purpose
• Ensure that the product is clinically safe to use
• Demonstrate that the product collects, stores and processes users’ information in a safe, fair and lawful way
• Be fair, transparent and accountable about what data is being used
• Be transparent about the limitations of the data used *
• Make security integral to the design and ensure that the product meets industry best practice security standards
• Ensure that the product meets all relevant regulatory requirements
• Ensure that the product makes the best possible use of open standards to ensure data quality and
interoperability
• Generate evidence that the product achieves clinical, social, economic or behavioural benefits
• Define the commercial strategy

23. A vision
The paradigm of the intelligent library
From searching to find a text to read
To interacting with the full text of the library collection
OR The living systematic review
Cox, Pinfield and Rutter (2019)

24. What would your vision of a
library powered by AI be?
Post your answers in the chat!

25. The impact of AI, machine learning, automation and robotics on the
information professions: A report for CILIP
https://www.cilip.org.uk/general/custom.asp?page=researchreport

26. Your questions and comments
are welcome!

27. References
• Cox A (2021) The impact of AI, machine learning, automation and robotics on the information professions: A report for CILIP
https://www.cilip.org.uk/general/custom.asp?page=researchreport
• Cox, A. M., Pinfield, S., & Rutter, S. (2019). The intelligent library: Thought leaders’ views on the likely impact of artificial intelligence on academic libraries.
Library Hi Tech. 37 (3) 418-435.
• Crawford, K. (2021) Atlas of AI. Yale University Press.
• Department of Health and Social Care. (2021) https://www.gov.uk/government/publications/code-of-conduct-for-data-driven-health-and-care-
technology/initial-code-of-conduct-for-data-driven-health-and-care-technology
• Glanville, J. (2020). Text mining for information specialists. In Craven and Leavey Systematic searchin:Practical ideas for improving results. Facet.
• Global Partnership on Artificial Intelligence (2020). Working group on the future of work, https://gpai.ai/projects/future-of-work/
• IFLA (2020) IFLA statement on libraries and artificial intelligence https://www.ifla.org/wp-
content/uploads/2019/05/assets/faife/ifla_statement_on_libraries_and_artificial_intelligence.pdf
• Jones, K. M., Briney, K. A., Goben, A., Salo, D., Asher, A., & Perry, M. R. (2020). A comprehensive primer to library learning analytics practices, initiatives,
and privacy issues. Jones, KML, Briney, KA, Goben, A., Salo, D., Asher, A., & Perry, MR, a Comprehensive Primer to Library Learning Analytics Practices,
Initiatives, and Privacy Issues. College & Research Libraries, 81(3), 570-591.
• McKinsey Global Institute. (2018). Skill Shift: Automation and the Future of the Workforce (Discussion Paper, May 2018). In McKinsey &Company (Issue
May). https://www.mckinsey.com/~/media/McKinsey/Featured Insights/Future of Organizations/Skill shift Automation and the future of the
workforce/MGI-Skill-Shift-Automation-and-future-of-the-workforce-May-2018.ashx
• Robertshaw, M. B., & Asher, A. (2019). Unethical numbers? A meta-analysis of library learning analytics studies. Library Trends, 68(1), 76-
101.
• Twidale, M. B., & Nichols, D. M. (2006). Computational sense: The role of technology in the education of digital librarians.
https://researchcommons.waikato.ac.nz/handle/10289/40
• World Economic Forum (WEF) (2020). The ten skills you need to thrive in the 4th Industrial revolution https://www.weforum.org/agenda/2016/01/the-
10-skills-you-need-to-thrive-in-the-fourth-industrial-revolution/