CILIP Libraries Rewired 2023 #3

A series of reflections from the CILIP Libraries Rewired 2023 event.

I was fortunate enough to receive a bursary to attend this first ever Libraries Rewired event in London on 10th November 2023. 

The aim was to connect “tech-curious colleagues”, suppliers and thought leaders on “all things digital.” The event was buzzing with enthusiasm and to get an idea of the atmosphere see the show reel on the  Libraries Rewired webpage. 

For me the most interesting and inspiring talk was the keynote address from Bill Thompson, journalist, commentator and technology critic.  

Suffering an AI Change 

“GenAI is about to do to librarians what Google did to GPs fifteen years ago, and you are not prepared”.

He dislikes the term AI preferring the phrase “computational rationality” created by Russell and Norvig in 2021. 

He suggested that there’s a role for librarians in explaining AI to the wider audience; it is not necessarily going to be come self-aware, it follows thought processes but not consciousness. However as we move into a world which runs on “computational power”  what is the impact on libraries, archives, information services and their contents? 

We need to consider: 

• what happens when searching becomes rewriting 

• how dependent will we become on computers to analyse data/content 

• how should libraries deal with content created by AI 

• how can libraries function among huge quantities of disinformation 

• the inherent bias in existing models 

• how to deal with inaccurate outputs 

• data centres using too much energy and water 

The historic role of libraries has been to navigate the world of information; we need to help people understand that information from AI sources may have been taken from various resource, rewritten, then repackaged. 

He asked what is a library, how can it succeed in this new climate?  

We must constantly make the argument for the thing that libraries do.  A library is the result of the decisions made by librarians, it is more than providing access to the entire mass of knowledge, it’s about 

• being the most empathetic element in the loop 

• having a human-shaped associative memory 

• being able to read emotions in a voice or a face 

• knowing what’s important and shaping the machine response towards it 

• being fallible like an organic being 

He left us with the message that  “It’s not safe to leave technology in the hands of capitalism. 

We should seize the means of computational reality!” 

His slides are available on his Wiki.

Books mentioned during the talk: 

The Atlas of AI by Kate Crawford 

Unmasking AI  by Joy Buolamwini 

The Coming Wave  by Mustafa Suleyman 

Artificial Intelligence  by Stuart Russell and Peter Norvig 

After listening to that talk I heard a discussion about The Future of Library Technology between Tom Shaw, Emma Adamson and Ken Chad. It covered engagement, UX, collaboration and integration.  Finishing with the quote from William Gibson:  “The future is already here but it’s unevenly distributed” 

I then heard from Sue Attewell, Mia Ridge and Daniel Van Strien on Harnessing AI and machine learning.  Mia discussed a project  from the BL working with the Alan Turing Institute using machine learning to identify people in census records and newspapers; it was a huge learning experience in upskilling staff and surfaced a lot of complexities. Daniel spoke about how libraries can leverage machine learning. He mentioned the Hugging Face hub which aims to democratise machine learning models.

Libraries could collaborate on a lot of projects; he felt it was important in improve information literacy in AI for library staff but also to offer training to library users. Students are already using AI tools. There were questions around how to upskill library staff and how to democratise it. The answer seems to be in collaboration and partnerships. 

In Do More with Data Luke Burton, Kate Lomax and Dave Rowe spoke about the importance of librarians being confident in using and exploiting data sources. Even in small projects using data can be influential.  Data knowledge skills are essential for advocacy; we need them to increase our understanding of our services and how to improve them. We can us data visualisation to tell stories, create maps and drill down to find out what’s happening and why. 

Ethics and Emerging Technologies was the final session of the day and brought together a panel of Nick Poole, Sue Attewell, Tom Shaw, Bill Thompson and Liz White.  First they addressed the question of avoiding biases; it was agreed this was a huge challenge. AI returns responses based of materials created at a time when there were more biased world views, how to avoid them? How to recalibrate responses from AI which contain those biases?

It was agreed that the role of libraries is important in this.  Another question was around transparency and authenticity. How can libraries provide guidance? Higher Education is focusing on the impact on teaching of student use of ChatGPT. There is the issue of privacy and helping people to understand their rights, the individual can feel excluded 

The final thoughts were around the importance of information skills for both librarians and the wider community. Librarians need to be part of the debate about where these technologies are going and how to support everyone.  But the debate must also be about what the role of librarians can and should be.