Computing for Humanity (Values in Computing)
Exploring human values in computing through an Artificial Intelligence focused workshop
Artificial Intelligence (AI) has seen rapid developments in the past twenty years. Concerns around the undesirable and unpredictable impact that AI may have on society are mounting. Leading AI thinkers and practitioners have drafted principles for “AI that would benefit humanity instead of causing it harm”. Underpinning these principles is the perceived importance for AI to be aligned to human values and promote the common good.
The aim of this 75-minute workshop is to help participants reflect, articulate and deliberate on the human values implications of computing on society by considering the impact of AI. We will do this through the use of tools and techniques developed as part of our Values in Computing (ViC) research at Lancaster University ( http://www.valuesincomputing.org ). ViC is about understanding how human values work in software production. We investigate both the easier to measure values such as the commercial success of a software product and the more difficult to capture values such as fairness, trust and public good. Our work helps to identify consensus and tensions within and between software development teams, stakeholders, end-users and society in general. By doing so, we may help anticipate technical decisions that could lead to breaches of moral standards and interfere with the very public good that software wishes to support.
There are three specific objectives to this BCS SPA workshop:
1. To give a shared high level understanding of AI
2. To give an understanding of ethics and human values, and how they may apply to AI specifically and computing in general
3. To deliver a group-based exercise to explore the articulation of human values in AI and computing
The aim is not to converge on a binding view of an AI manifesto (we don't have the time!), but to lead a debate, create awareness on knowledge gaps, skills and resources within the group, and to foster consensus on what attendees agree and agree to disagree on.
Anyone can attend – no previous knowledge of AI or Ethics/Values studies required. Background check on attendees though – anyone working in AI or on computer ethics and values make yourself known!
Benefits of participating
An experimental, provocative and informative session to explore individual and group values in relation to AI.
Exploring the use of our “Values Q-Sort” cards and tools. After the workshop - to be more human values aware in your teams, projects and organisations. To keep in touch with us, and potentially research further with us…
Slides, sets of AI Values Q-Sort Cards and tools.
We’ll need OHP, flip charts for each team, post-its and pens, and to be able to stick things to walls.
1. Introduction – our work on values in computing
2. AI Baseline – shared understanding of AI (use and scope in this workshop)
3. Personal Portrait
4. AI Values Q-sort guided exercise - 2 iterations towards an “AI Manifesto”. Formation of 4 groups.
4a. AI V-QS activity and cards introduced. Prompt for what AI systems should and should not do or be. Ask for missing statements or values. NOTE: definition of AI left open - each group can question/argue/decide.
OUTCOME: 4 group V-QS grids, ask to take photo when Q-sort reasonably stable; identification of top/bottom 5 values and any missing statements.
4b. Merge 4 teams into 2 teams; ask each to compare and contrast their Q-sorts and top / bottom values.
OUTCOME: group note down their 5 agreed top/bottom values and any missing statements;
4c. Teams report back on their key values, principles, agreements, disagreements, tensions, and missing points.
OUTCOME: Team flip chart = first steps towards an AI Manifesto.
5. Review of the tools we used and how they were designed.
00:00 – 00:10 Values in Computing and AI - introduction
00:10 – 00:15 Personal Portrait – your profession, motivation and reflections on AI
00:15 – 00:60 AI Values Q-Sort Exercises leading to the co-creation of the group’s AI Manifesto
- Q-Sort intro (5 mins)
- Q-Sort 1st iteration (15 mins)
- Q-Sort 2nd iteration (15 mins)
- Report back and debrief (10+ mins)
00:60 – 00:75 Explanatory slides: motivation and explanation of tools used. Wrap up, what will you take-away?
Draft manifestos on a wall at SPA for others to see.
Report to the ViC website – the slides we presented, photographs taken, summary of the session including your draft manifestos.
Optionally: Poster or paper at a future conferences (if participants consent to us using their output).
An invite to keep in touch and work with us in the future.