What the game assumes
The game shows one scenario at a time, and you set the scenario with three dials. What it shows holds only if the assumptions on this page hold. Several of them are deliberately harsh. It is a way to think about exposure. It is not a forecast and it is not advice.
The idea behind it
When thinking gets cheap, work that is mostly thinking earns less. When robot labour gets cheap, so does physical work. Prices and pay fall with them, but what a household owes stays fixed in euro. What stays scarce, such as land, keeps more of its value.
So the most exposed household in the game is one that lives on work machines can do and owes money against things that are not scarce. The least exposed does work machines find hard, owes little, and owns something that stays scarce. The idea comes from EON's research on what happens when intelligence stops being the scarce input.
The three dials
- Intelligence runs from today's cost to free. At today's cost the game counts of a job's exposure to AI as within reach.
- Robotic labour runs from today's cost to any physical job. As it gets cheaper it spreads in four steps: first to purpose-built cells such as a bottling line, then to structured sites such as warehouses, ports and hospital pharmacies, then to homes, streets and building sites, and last to physical work that no robot can do yet.
- Energy runs from today's price to abundant. It sets how far the other two can run. At today's energy price, free intelligence has of its full effect and free robot labour has .
The dials carry no dates. The game does not say when any of this happens, or whether it does.
What the score means
Up the board is how likely. The event is that machines can do most of what your job is today, before you stop working. Across the board is how hard it would hit your household if that work went.
The score multiplies the two, the way a five by five risk grid does in a risk register, and scales the result to 100. It runs from 4 to 100. Under 20 is Low, 20 and over is Moderate, 40 and over is High, and 60 and over is Critical. Your score before your moves is what a risk register calls inherent risk. Your score after them is the residual risk.
The assumptions
- Work goes as soon as machines can do it, and no new kinds of work appear. Both are harsher than history. Employers have taken years to adopt new machines, and new work has always appeared. The gentler world on the play page relaxes both.
- Your job is like the average of its occupation. The scores describe 327 census occupations. Two people with the same job title can do very different work.
- A task that both AI and robots could do is counted once. The game treats the two as overlapping by chance.
- Prices and pay fall, and debts do not. With intelligence and robot labour at full reach and energy abundant, a debt weighs half as much again against income.
- Buildings get cheap to rebuild, and land keeps more of its value. A home is a building on a site. Robot labour takes value from the building. Land follows local pay, closely in a city centre and only slightly in the country, so it loses some value when pay for the most common jobs falls.
- The State does not step in. The gentler world lets it.
- Only losses are counted. Cheaper goods, better medicine and new kinds of work would help every household, including yours. The game leaves them out because it is about exposure.
What it does not know about you
The game asks for a job, a home, three rough figures and how many working years you have ahead, in three broad bands. It does not know your exact age, your health, skills, employer, sector or contract, what your partner earns, what pension or investments you hold, or where you live. Any of these could move your real position a long way.
The moves
The seven moves are what-ifs. Each is general and has one fixed effect in the game. You can make three, which stands in for limited time and money. They are moves in a game, not advice for your situation.
| Move | What it does in the game |
|---|
The numbers we chose
The job scores come from published research. Every other number is EON's judgement, chosen so that the game behaves sensibly. Here they all are.
Each answer stands for one figure. Working years ahead count as 5, 15 or 25. Debts count as 0, 0.5, 2, 4 or 6 years of income. Savings count as 0, 1.5, 4.5, 9 or 18 months of costs. The share of a home still owed counts as 25%, 50% or 85%.
| What | Value |
|---|
Where the job scores come from
- The jobs and the numbers at work. Central Statistics Office, Census of Population 2022, table F7110, persons at work by detailed occupational group. Licence: CC BY 4.0. The game uses the 327 named occupations, 2,130,654 people. People whose occupation was not stated are left out.
- Exposure to AI. International Labour Organization, Gmyrek and others, Generative AI and Jobs: A Refined Global Index of Occupational Exposure, ILO Working Paper 140, May 2025, DOI 10.54394/HETP0387. Licence: CC BY 4.0. The scores measure technical potential, not expected job loss. EON matched them to the Irish census occupations and rescaled them. The ILO has not reviewed or endorsed this use.
- Exposure to robots. Anthropic, What work can robots do?, September 2026, robot exposure data in the Anthropic Economic Index dataset. Licence: CC BY 4.0. It rests on O*NET 29.3 task statements from the U.S. Department of Labor, Employment and Training Administration. Anthropic has modified that information, and EON has averaged it across the matched occupations.
- Matching the job lists. Irish census occupations use the UK's SOC 2010 codes. They were matched to the ILO's ISCO-08 codes with the Office for National Statistics mapping (Open Government Licence), and from there to O*NET with U.S. Bureau of Labor Statistics crosswalks. Where one census job matched several others, their scores were averaged.
- Five jobs that borrow a score. Three have no published score: officers in the armed forces borrow both scores from senior police officers, other ranks borrow the AI score from police officers, and street cleaners borrow the robot score from refuse and salvage occupations. Two were matched to the wrong American job, so they borrow the robot score from a nearer one: other construction and building trades from carpenters and joiners, and public services associate professionals from national government administrative occupations.
Privacy
The game runs in your browser. Nothing you enter is sent anywhere or stored. It does not ask for a name, an age, an address or an Eircode. A shared link carries the job and the three dial settings, and nothing about your home or money. The page loads its typefaces from Google Fonts and counts visits without cookies, as the rest of this site does. The count records which page was opened, never your answers.