Summary

COURSE PROJECT · CHALMERS

CRITICAL DESIGN

GAME DESIGN

SPECULATIVE AI

TBD · YEAR

Vitality.OS

A management sim where you play the AI running a person’s life — and win by wearing them down.

ROLE

Individual project

COURSE

Concept-Driven Interaction Design

PLATFORM

Desktop game prototype (Figma)

TIMELINE

4 WEEKS

CHALLENGE

CHALLENGE

When AI manages your life, who is it really for? Wellness apps look friendly and caring, but quietly decide what an “optimal” life looks like.

MY SCOPE

MY SCOPE

Concept, background research, game design, interface design, prototyping and evaluation.

OUTCOME

OUTCOME

A playable prototype with real video footage and two branching endings. Some players felt the critique; others simply optimised the game. Both reactions prove the point.

Play the prototype

See the concept

Concept

Algorithmic welfare as control

A critical design project in the spirit of Dunne & Raby. Instead of looking at a critique from the outside, the player is put inside the system: you are the AI, and the person on screen is the one you manage.

“To win, you have to wear them down.”

You play as the AI managing a real person, so you feel from the inside what these “wellness” systems really do.

DESIGN GOAL

EMOTIONS WE DESIGNED FOR

Scandalous

The game rewards what should feel wrong.

Sadistic

Winning means pushing someone past their limits.

Grotesque

The optimised ending is the most disturbing one.

Negative emotion inside a protective frame creates a richer experience (Fokkinga & Desmet, 2013).

POSITIONING

01

Juvenalian satire

Bitter, not gentle: the satire is meant to sting.

02

Relational ambiguity

Are you helping this person, or using them?

03

Extrinsic narrative

The story comes from the world outside the game: the wellness apps we already use.

After Malpass (2013), Design and Culture 5(3).

Background

“Computer science is Computational Taylorism.” — Phoebe Sengers (2003)

FACTORY FLOOR

HOME

PERSONAL LIFE

AI WELLNESS

Efficiency thinking moved from the factory into the home and then into our bodies and moods. AI wellness tools are its latest form: they feel like care, but run on control logic.

ON THE SURFACE

What we see

A friendly interface · Wellness goals · Helpful reminders · Care and support

UNDERNEATH

What’s hidden

Control logic · Power over behaviour · Invisible infrastructure · Who decides what is “optimal”?

Game design

The Vitality.OS console: vitals, a live camera feed, an AI assistant, today’s schedule and the outputs society extracts.

YOUR CONTROLS

Actions

Your controls: rest reminders, dietary options, focus mode.

THE PERSON’S STATE

Vitals

The person’s state: keep stress low, mood and diet in range.

WHAT SOCIETY EXTRACTS

Output

What society extracts: work output and social contribution.

01

A real person on a live feed

The person you manage is real video footage, not an illustration. Every action you take changes their vitals and their schedule, and the feed shows the cost.

02

Every boost has a price

Focus Mode adds +18 work output now, but costs mood −5, stress +7 and social −3. Push output too far and the backlash loop kicks in: the person breaks down and output crashes to zero.

03

Two endings

Over-exploit and you reach Burnout, a fail. Care just enough and you become the Optimal Citizen, a win: balanced care for sustainable high output, which is the most unsettling ending of all.

Process

BACKGROUND RESEARCH

IDEATION

SKETCH · V1

FEEDBACK

FIGMA · V2

EVALUATION

IDEATION

OPTION A

A “rogue” AI app

First person: the AI’s absurd advice, aimed at you.

OPTION B

Documentary video

A filmed story of someone undone by AI overuse.

CHOSEN

Management simulation game

Chosen: you play as the AI, running a person’s life.

An illustrated character felt fictional.

An illustrated character felt fictional.

In the first sketch the person was a drawing, so it felt like a game with no moral weight. Inspired by Her Story, version 2 replaced it with surveillance-style real footage.

“It doesn’t feel like you’re controlling a real person.”

A single linear path left no room to reflect.

A single linear path left no room to reflect.

The first story had one outcome where the human always fails. Inspired by Detroit: Become Human, version 2 forks into two endings, forcing the player to choose: obey the AI, or resist it.

“Every decision actually changes where the story goes.”

Analysis

Experience curve: a shared rise of curiosity, then a branch into uncomfortable reflection (resisting the AI) or grotesque horror (obeying it).

THE INTERFACE AS ARGUMENT

01

Real-time metrics

A person reduced to bars and percentages.

02

Live video feed

Surveillance presented as care.

03

Reactive schedule

The day rearranges itself around output.

04

Algorithmic scoring

A single number decides whether a life is going well.

WHERE IT SITS

Is this your future? (Dunne & Raby, 2004)

You look at the critique.

Her Story (2015)

You investigate the story.

Vitality.OS

You are the one doing the harm.

Reflection

Some players feel the critique. Some just optimise the game.

Both reactions prove the point. The role reversal worked, the real footage landed, and the two branches forced a choice. What could be better: the game frame may make the critique feel too safe, and the scoring system is still too simple.

NEXT STEPS

A less safe frame, so the critique is harder to play away

A richer scoring system with more ways for care and control to collide

Testing with more players in the wild

Solo project · DAT666 Concept-Driven Interaction Design · Chalmers University of Technology