Paths
Eleven decisions between 2026 and 2037. Choose in each one and see which future your path leads to, starting from where we are today.
How it works and how we label
The path starts at the today point on the matrix. Each option moves the point towards more or less peace (how much we regulate and cooperate) and more or less frontier (how fast AI advances). At the end, the path leads to the nearest future, with your exact position. If the point is already at the edge, the effect is clipped: today the frontier is already at its maximum, so options that speed it up do not move it further.
The effects are our estimates, not model outputs. The futures model takes only two numbers. How much each decision moves them is an editorial choice, and we show it next to each option so it can be debated.
How we label. We rate each option on four criteria: whether it reduces or increases harm to others, whether it has the consent of those affected, whether it spreads or concentrates benefits and power, and whether it can be undone. Each criterion adds +1, 0 or −1. With 2 or more, the option is ethical; with −2 or less, unethical; in between, ambiguous. The rule is fixed; the ratings are ours and debatable.
What follows from these estimates. If you decide nothing, you stay in Race, where we are today, and most possible combinations end there: leaving takes several decisions in the same direction. If you always choose the ethical option, you reach Pause, because the criteria reward consent and reversibility, which tend to come with more rules and less speed. An ethical option does not necessarily lead to the best future, nor the other way round.