Exploring potential AI futures: Agenda, templates & learnings from my first Speculative Design workshop 🔮
Intrigued with the method, last week I ran an experiment with my 1st attempt at facilitating a speculative design workshop. Struggling to find practical resources online during prep, I decided to share my workshop agenda & reflections here for all those looking to run an experiment themselves.
What is speculative design?
Speculative design is a practice to spark thoughts and discussion about the social, cultural, and ethical implications of emerging technologies. It’s about developing potential future scenarios, anticipating consequences, expressing interesting tensions in tangible artefacts, and ultimately guide us in designing preferable futures.
A pretty new field, similar practices go under the names of critical design and design fiction (while its originators try to hash out the differences, if any) and the methods are as ambiguous as the terms. Regardless, the results stand their ground. Equal parts strategy and art, it is used within companies to inform competitive business strategies as well as to guide ethical implementation of new products.
If you’re curious how the pros do it, some studios doing great work in this space are:
- SuperFlux (translating future uncertainty into present-day choices)
- Extrapolation Factory (experimental methods for collaboratively prototyping, experiencing and impacting future scenarios)
- Near Future Laboratory (to understand how imaginations and hypotheses become materialized to bring the present into new, more habitable near future worlds).
If you want a bit more background, I recommend reading this post: Can Speculative Design make UX better?
Disclaimer: I am far from an expert and do not guarantee that what you’re reading here is a ‘good’ representation of the speculative design practice. I’m just learning, experimenting, and sharing my experience.
Why I wanted to test it in my project
My thesis project explores the opportunities & challenges of using ML (Machine Learning)/AI (Artificial Intelligence) to create meaningful user experiences.
These new tech applications (such as self-driving cars or emotionally-aware services) are cool and have obvious value, but raise a lot of questions around privacy, agency, inclusivity, dependency, and more.
Interested in exploring these experiential, societal and ethical consequences and how to design helpful interactions with algorithmic systems, I decided to test speculative design as a method for exploring these collectively.
I shared the following invite on MOBGEN’s Slack channel and my Instagram and within no time, 10 participants signed up.
Why this post?
During the prep for this workshop, I really struggled with how little practical resources I could find online. While many practitioners shared triggers and artefacts, I couldn’t find a single source that shared their process, exercises, or workshop agenda. Being the change I’d like to see — I’ll share everything including the materials and learnings of my experiment with you here. A post on outcomes might follow, this one is about process.
The full workshop agenda & learnings step by step
I’ll walk you through the workshop agenda including exercises, and reflect on what worked (+), what didn’t (-), and how to improve next time (>). You can also access the full deck here and A3 templates to print here.
The agenda looked a little something like this:
Pick your own prompt
To start off, I invited the participants to construct their own prompt consisting of a tech trend (one of 6 AI capabilities I explained in the introduction), and an industry/life area (free choice). The timespan was set to 2030, 11 years into the future.
+ Allowing to pick their own industry made it more relevant and directly applicable to their work.
- Allowing to pick their own tech trend took more time than necessary because of disagreements within groups.
> Alternatively, let each individual choose for themselves (e.g. by hanging the 6 options on walls around the room and asking people to stand with their favorite one) and let the groups form around that.
World building 🌍
Once decided, I asked them to fill out the following template.
+ Giving a structure like this proved useful and was necessary because some elements are used for the force fit in the next step.
- I tried in 15 minutes and it was too short so 20–30 would be better.
> Ideally, you would leave a bit more space for free discussion among the participants and maybe an unstructured mindmap in the middle in addition to the predefined elements.
To begin imagining potential future scenarios in their world, I resorted to the classic force fit exercise.
+ Ideas were generated and sufficient to continue with. In this step it’s okay for ideas to be cliche. 4 per person is more than enough.
- During the workshop I took 2 industry elements and then added ‘using agentive technology’ which did not work very well as 1) participants don’t have a full grasp of the tech trend and 2) stuffing 3 elements together is a bit much and they lost context.
> Next time, I would mix the quintessential scenarios/user jobs with the new tech prompts. In the previous step, the tech trend must be broken down into abstract elements or first principles that work well as prompts, for example describing capabilities or case studies.
Consequence wheel 🔮
The consequence wheel is a well-known tool and proved useful as a framework to think and discuss through first, secondary, and tertiary consequences.
+ Participants really resonated with format. Interesting discussions emerged
- I gave them 15 and it felt rushed. 30 minutes would be better. Some threads were based on circular reasoning and lacked specificity or depth.
> This might be helped by showing some ‘good’ and ‘bad’ examples, or even building out a few threads collectively to show the types of questions to ask.
Mapping tensions (in future user needs)
In the next step, I asked the participants to map tensions in future needs based on observations, questions and concerns around their potential future experiences. Another prompt can be around hopes & fears or best & worst case visions of this future.
+ Capturing these tensions comes natural as they’ve already come up in the conversations before.
> It would be interesting to go more into this. What is your definition of the word you’ve decided to use? Where on the spectrum do you see your preferred future? In which situations would the more extremes be helpful?
Based on the work so far, I ask the groups to pick one story-consequence-future user need threat they are most intrigued by and summarize their POV before we move into familiar ideation phase.
> I didn’t have this step before the workshop and added it on the spot. The user needs as they exist in the outer layer of the consequence wheel are not clearly defined enough. Starting from a good POV statement is crucial for successful ideation.
Ideation as we know it.
+ People already have ideas from before and find idea generation easy.
- Ideas are sometimes cliche and mostly service-based.
> Probably with more prompting and guided ideation (even a round of building on each-others ideas, or force-fitting with other elements), I trust participants would think of other artefacts such as products, jobs, spaces, agents, and interactions beyond the obvious.
+ People have interesting ideas.
- But pick the more cliche ones to move forward with.
> Could be helped by continued discussion or dot voting across the groups to bring out the most interesting concepts.
Prototyping future artefacts 🎨
+ People were motivated to prototype and build their artefacts.
- I failed hard at time management and unfortunately didn’t have space to make decent prototypes.
> Plan at least 30 minutes for this.
Collecting user feedback
> Ideally, the groups would take their artefact, present it to random people around (outside of the workshop) and collect their first thoughts and feedback.
+ People like presenting their ideas and how they arrived there. Also gives the different groups an opportunity to re-connect and learn from eachother.
- People are quick to go into ‘pitch-mode’ rather than presenting the tensions and complexities of their concept.
> A more structured format to present their process might be helpful.
Reflections & Outro
It’s important to bring the relevance of this exercise back into the participants’ day-to-day reality.
+ A few minutes of free discussion was enough to bring about an interesting discussion. I asked how the tensions we mapped might already show up in their work today, and how thinking about potential futures might inform (design) decisions?
- Some do struggle to see the applied value of speculation or how and when to integrate it into their design process.
Yes, speculative design is valuable because:
- Everybody holds visions and the ability to form visions about the future but rarely gets the opportunity to express, probe, and develop them.
- As technology is becoming increasingly important and consequences increasingly complex, speculation might be our only way of anticipating.
- Thinking and talking about the future, even in a speculative way, brings home the fact that we are indeed shaping it everyday.
- Bringing back that awareness into our day to day work, we can all play our part in building futures we wish to inhabit.
Yes, facilitating speculative design is challenging because:
- While interesting angles naturally arise in conversation, your process needs to somehow capture and build out these nuggets.
- To steer the conversation and make sure someone takes note of the interesting nuggets, you might want a facilitation style that’s more participatory. It is possible without, but it might help to have 1 facilitator/moderater per group.
- Going beyond the cliche black-mirror-esque visions requires guidance and probing.
- Not everybody sees the value of this ‘wavey’ stuff. That’s just a challenge for us to ensure it creates value, and to articulate its value.
All in all, I’m very happy with the learnings. The workshop itself wasn’t all that but it was a great reminder that it’s okay to try something new and run experiments. That said, the participants learned new things and enjoyed themselves, so did I. Interesting angles on future user experiences with algorithmic systems were explored, discussed, and developed (which feeds into my thesis and I might share more on later).
If speculative design is something you’re interested in, I recommend you to run an experiment yourself (and ideally, publish your reflections in a Medium post). It’s the best way to learn.
If you enjoyed this read, please share. If you are a fellow practitioner, please connect. If you are human, please have a beautiful day.