Who will be responsible for decisions AI suggests? Four answers from four worlds of the Future Decoder method. They come from participants of the scenario games, not from experts.
Autopilot world (CS)
the system decides, the result is not yours
Nobody is responsible; the algorithm decides.
- How it develops: The state lacks resources, so it automates collection and redistribution. The corporation or the state is the subject; everyone else is an object. Systems dismiss people by efficiency, and it’s unclear how to adjust what you do.
- The price: Psychological resilience wobbles. The system is dumb and controls simple things; black markets and stagnation appear.
- Signal: Systems dismiss employees based on efficiency metrics.
- Heard at the game: “Responsibility lies with the algorithm”
- Step for this week: When you get a decision from a system, ask: who exactly stands behind it?
Rules world (CE)
you take part by shared rules
Passed an AI exam? You’re responsible yourself.
- How it develops: AI in clinics sorts requests and makes preliminary diagnoses. Like a car: you’re behind the wheel, so it’s your fault. The state introduces an exam for using AI: pass it and you take responsibility. Parents take an exam so their child can study with a neural network.
- The price: The exam is insurance for the system, not a service. Responsibility is increasingly shifted to the person: “you didn’t take care of it yourself”.
- Signal: A doctor is personally responsible because of their qualifications; with AI it’s unclear who is (a participant’s experience; his wife is a doctor).
- Heard at the game: “You were warned, so check”
- Step for this week: Before an important decision based on AI advice, check it as if you were signing it with your own name.
Autonomy world (DS)
Nobody has designed this world for this question yet.
Community world (DE)
you create together with others
Problems are solved at the level where they arise.
- How it develops: The state is unnoticeable, like an automatic payment: people organise work, study and health themselves and turn to it in exceptional cases. A problem goes higher only if it can’t be solved below. Institutions are temporary: task solved, they dissolve.
- The price: Who controls the controllers? Order may rest on a hidden force.
- Signal: Elinor Ostrom (Nobel Prize): shared resources can be managed by a community without control from above.
- Heard at the game: “If I can change the brake pads myself, nobody needs to inspect”
- Step for this week: Pick one problem you pass “upstairs”, and try to solve it yourself or with neighbours.
Source: “AI and the Human” cycle: game of 11 Jul (Saturday); “AI and the Human” cycle: second part, 28 Jul (a consultant aged 40). Draft of the Future Decoder knowledge base, updated after every game. All questions