The Way Out Must Remain Open
Preparing for a technological future can make that future harder to reconsider. When applicants, employers and universities adapt to what they expect others will soon require, each decision may be reasonable. Together, however, they can create dependencies that leave later, better-grounded judgement without practical effect. This paper examines that mechanism, collective anticipatory self-binding, through hypothetical models of AI adoption, archaeological evidence from the Fremont complex and research on human–automation interaction. It asks when the loss of opportunities for correction requires justification, and what keeps technological commitments revisable. Drawing on published sources, it distinguishes the formal possibility of reviewing a decision from the practical capacity to change it. It then develops five dimensions for assessing revisability: scope, uncertainty, possible adverse consequences, substitutability and the power to secure correction.
