AI Ethics and Principles at State Library Victoria – Research, process and analysis

Research methods – What we did

This research sought to frame how the personal motivations of staff map onto their current uses of AI. This section shows how we developed the ideas put forward in this report.

Soft ethics provided the inspiration for our methods. In the absence of formal policy and guidelines, we wanted to explore how people are approaching their AI use through their encounters with the situation at hand, through a matrix of their values, organisational systems and interpersonal relationships.

We used a mix of group workshops that emphasized social connection, and one-to-one conversations and interviews, which supported staff to share thoughts and ideas anonymously. Staff were invited to the workshops using an EOI process. We ran 7 sessions in total, 6 for mixed groups of staff and one for the executive team. Each workshop included 4-23 participants. The same workshop process was replicated across all sessions with staff and executive.

We selected our 14 interviewees by approaching people during the workshops. We also invited several individuals who had not self-selected through the EOI process, to achieve a balance of representation across directorates, junior and senior positions, and the executive team. The interviews allowed us to capture detail about people's fears and hopes for AI innovation in an intimate, confidential setting.

Workshop process

We began with a system map that placed uses of AI within the context of internal and external relationships

Step 1: System map

Participants were invited to create a mind map that framed their use of technologies, responding to the following points.

•            Name an AI technology that you use in your day-to-day library practice

•            What task are you using this technology to perform?

•            Who is involved? Who is the outcome shared with?

•            Map all the technologies that you engage with, or as many as you can think of.

Step 2: Personal values

We invited participants to shift into a more reflexive mode and consider the personal values that motivate them in their role. Values are one of the factors that influence personal and collective ethics. We asked people to respond to the prompt:

Can you name a non-negotiable personal value that you bring to your role within the library, or that guides your library practice? More simply: what drives you to do your job?

Step 3: Benefits, challenges and values

In step three, we asked participants to pick a single technology from the map created in step 1 and consider the benefits and challenges that this technology creates for their library practice. We invited staff to use their list of benefits and challenges to consider the following prompts:

Is there anything in your use of this technology that compromises your non-negotiable value?

Are there specific ways that this too supports you in achieving your non-negotiable value?

For example, does using chat GPT assist your goal of making high quality information more accessible? Or does automating meeting notes using co-pilot support your goal to work collaboratively with colleagues? Most participants concluded that their values were both supported and undermined by the introduction of AI, giving a nuanced perspective on where values and technology meet. Our aim was to move away from simplistic good/bad binaries.

Step 4: What do you hope for?

We closed the workshop by asking participants to consider the following:

What is the most concerning thing the library could do?

What do you hope for?

These questions raised the stakes around personal values to provoke consideration of how participants would like the library to operate into the future. Concerns and hopes are future focused expressions of values. They further illuminate how individuals see the library’s cultural, social and moral role in society.

Analysis

• We met to debrief initial impressions from the workshops
• We listed the themes reoccurring in steps 3 and 4 to identify shared values that repeatedly occurred for staff
• We considered points of ethical tension and opposition, and pain points where AI caused people to diverge from what they value at the library
• We used responses to Step 4 to link people's personal values to broader hopes and concerns for the future of the library.
• We attributed equal importance to descriptions of incidental and unintentional and AI uses, where participants knew that they were using AI but struggled to name specific technologies
• We cross referenced workshop data with interview transcripts to illuminate more extreme positions. The workshops reflected nuances as people with different perspectives responded to each other. Interviews sometimes reflected more intense positions, for example, a trades union member who opposed AI for multiple ethical reasons but was particularly concerned about AI enhancing existing workplace inequalities.
• Where people reported actively avoiding AI, we took this as a form of engagement, feeling that AI avoidance is a specific ethical position with particular consequences for individuals in the workplace.

Email us at [email protected] to request a copy of the report. Please include “AI Ethics and Principles at State Library Victoria” in the subject line