SenseCraft Game Design

A game platform to build citizen's capacity to engage with complex evidence

​

Jack Park�jackpark@topicquests.org

Marc-Antoine Parent�maparent@conversence.com

https://creativecommons.org/licenses/by-nc/2.0/

© 2023 TopicQuests Foundation and Conversence

Photo by Johannes Plenio on Unsplash

​

Collective intelligence and citizen participation

• Every public decision should be justified
• Goals, evidence, provenance, context, limitations…
• vs opaque decisions processes: Bureaucracy, Hidden algorithms, LLM
• Small diverse groups make better decisions on complex issues
• Self-selection bias, sampling bias
• Group dynamics do not scale
• Size scaling: Time to hear everyone out
• Complexity scaling: Simplify issues (vote) or take time to master background knowledge
• Divide in workgroups: divided understanding
• Language is sequential, maps are fractal
• Compositionality of units allows more contributions and contributors

How to engage citizens with complexity

• Map literacy is a skill
• Good map writers help in medium assemblies
• Asynchronous crowdsourcing leads to coherence issues
• How to motivate citizens to learn to engage with complex maps?
• People spend time to master games
• Even with very complex rules!
• Games provide a context with a common goal

Early gamification: Foresight Engine, MMOWGLI

• Individual play
• Conversation tree structure
• Points for provoking responses
• Moderator bonus
• Participant bonus

​

​

Note: Current state of our prototype uses a similar conversation structure

Early gamification: induced gaming dynamics

• Dynamics induced by “playing the rules”
• Reactive play (fast thinking)
• Competitive engagement can dominate inquisitiveness
• Moves to establish personality
• Contributes to community, not to structure. Does not scale well!
• Lower signal to noise ratio
• Not worth the (high) cost of game preparation and moderation
• Tree structure does not encourage convergence

Jack Park was asked (2010): �“How can we have civil conversations online about politics?”

World of Warcraft meets global sensemaking

“I would rather hire a high-level World of Warcraft player than an MBA from Harvard” –John Seely Brown (2012)

​

• Social side of the interaction will mostly happen within-team
• Team will want to keep disruptive behaviour in check to protect reputation
• Roles lead to specialization and reduce context switching. Enforced diversity
• Roles and aliases allow disengaging from one’s point of view
• Mentorship dynamics within the team

Design for deep listening and nuance vs reactivity

• Turn-based vs Real-time
• Each team presents positions text
• Each team maps other team’s scenarios
• Map unification and clarification
• Reward exhaustive mapping
• Reward teams that contributed an element (divided among teams: rewards originality)
• Reward teams that mapped that element: rewards attention to detail
• The map element has to be anchored in the other team’s presentation
• Encourage clarification questions
• vs False consensus, Dunning-Kruger
• Provide alternative plausible interpretations
• Verify plausibility of alternative with the other teams

Design for coopetition vs adversarial dynamics

• Mutual scoring
• Team must spend tokens to identify quality contributions of other teams
• Reward synthesis that bridge elements of previous proposals
• The contributions must be recognized
• The integrated elements will also count towards the original contributor’s total
• Track adoption of proposal and synthesis by teams, favour change (Change My Mind)

Design for epistemic curiosity vs playing fixed rules

• External evidence as minimal justification
• Demonstrate understanding of the evidence (ask for methodological limitations?)
• Can be prone to adversarial dynamics
• Contested status for contradictory evidence
• Resolving contested evidence with meta-reflexion
• Identify shared heuristics (criteria) that would allow to decide contested truth (value)
• Argumentation schemes
• Look for blind spots, bias
• Encourage asking the why of the why
• Deep democracy

Co-design as a meta-goal

• Engelbart’s vision: Coevolution of humans, tools and processes
• Expand to ecosystem
• Players define Quality
• Mutual scoring as first step
• Risk of collusion?
• Identification of cognitive patterns
• Player-enriched badge system
• Building a library of meta-heuristics
• Eventually: co-designed game dynamics?
• Players design their own team dynamics (through roles)
• Defining what pro-social behaviours the game should encourage
• Construct a grammar and representation of meta-reasoning

How would we include AI in the process?

• AI as research assistants
• Players will use AI. Accept it, but red flag for hallucinated evidence.
• AI as game narrator
• Appropriate for the big picture, game play highlights
• AI as sparring partner
• Learn to guess the AI-simulated team!
• Learn to identify bias or unverified assumptions in AI-generated positions
• AI as reference point of consensus reality
• When a team makes a claim that AI did not suggest…
• It could be extraordinary original!
• But it probably requires extraordinary evidence
• AI as design partner

Co-design with the Collective Intelligence Community

• A lot of Collective Intelligence sub-communities working on their own
• We need to exchange tools, data, methodology
• Co-evolving a tool ecosystem
• We need a shared protocol for CI interoperability
• My previous contribution: Catalyst Interchange Format (2014)
• Co-design the requirements!
• Anything can be re-interpreted.
• Deep recursivity
• Towards HyperKnowledge
• https://sensecraft.garden
• https://hyperknowledge.org