SmartThing Generator

The SmartThing Generator is somewhere in between being a game and an ideation exercise. It’s a future-building activity that uses a cyclical pipeline of three groups and strategy decks to construct the narrative. The groups check and balance one another to illuminate pitfalls in A.I. that are easy to miss and/or convenient to ignore.

Instructions for Use

To Start:

The three groups each start the game with a set amount of resources. These resources are broken down in to five types of “currency.”


Consumers and Designers each start with five.

*I may actually get rid of this form of currency - I’m not sure it adds anything meaningful to the process.





Consumers and Appraisers each start with five of each.

Group 1: Consumers

My main objective in this project was to examine fundamental themes in design for A.I., not to make a system for world-building; so, rather than creating entirely new content for this group’s role, I decided to modify a game that we’d researched called The Thing From the Future. The consumer group starts things off to establish a context. Group 1 draws from Thing cards arc, terrain, and mood. Using these they invent a backstory for the group of consumers they’ll represent. From this context, they’ll describe some characteristics of an A.I. that would most benefit them.

Group 2: Designers

Using Group 1’s specifications, Group 2 uses ideation cards to build an A.I. proposal. Cards describe the basic quality of the input/output of the A.I. and the processes used on them. They are arranged into a sort of formulaic abbreviation for the designed thing. Cards are supplemented with an illustrated prototype of the thing and how it works.

Writing is allowed in moderation. Group 2 presents their proposal to groups 1 and 3. Walk through technological process(es) of the thing, specify the data/content used for the input and generated for the output.

Group 3: Appraisers

The appraisers review the design proposal and make a subjective determination about how this smart thing (including its manufacturing) will affect the consumers. For features that are determined to disadvantage the consumers, the consumers will “pay” the with the correlating resource. For features that are determined to benefit consumers, the appraisers will pay the consumers. Lastly, the appraisers will update the consumer context according to their appraisal of how the smart thing has affected the consumers.


The cycle repeats with the consumers describing new wants and needs based on their standing after the first round in relation to both the context and their resources.

The game is cooperative. If any team runs out of all its resources, everyone has lost the game.

Materials Used:

Project Brief:

Strange New Creatures Design Research Project, Phil van Allen

AI systems are inherently strange because of their mix of human parentage, algorithmic quirks, and savant syndrome with “islands of genius.” Our research project will embrace this odd mix in a series of design experiments that strive to create AI creatures with form and behavior that are “natural” to this strange inner character.