Who should decide the limits of a technology
A technology does not set its own limits
A system can work exactly as designed and still raise questions about where it should be used. A tool that identifies people, for example, may be technically effective yet have different consequences on a personal device, in a school or in a public square. Knowing how to build a technology does not independently settle which applications are acceptable. That decision involves evidence, rights, collective objectives and responsibility for mistakes.
Anticipating problems before widespread adoption
The OECD's 2024 anticipatory governance framework brings together values, assessment of possible futures, public engagement, adaptive regulation and international cooperation. Its aim is to discuss consequences during development, before choices become difficult to change. It provides governance guidance, not automatic authorization or a universal law covering every technology.
Technical knowledge and legitimacy are different
Specialists can explain error rates, operating conditions and system limitations. This information is necessary, but does not independently decide who should bear the risks. Affected people may identify consequences that technical tests did not examine. Authorities must justify decisions within their responsibilities. Companies must answer for what they offer. Dividing these roles does not mean that every opinion carries equal weight on every question; it means identifying which kind of knowledge or authority each decision requires.
A hypothetical example in a school
Imagine a school considering an automated attendance system. The supplier reports high accuracy in its tests. Before purchasing it, the community could ask which problem needs solving, whether a less intrusive alternative exists, what data would be stored and who could access it. Errors, challenges and arrangements for people unable to use the system would also need discussion. This example does not describe a specific legal requirement. It illustrates useful questions for assessing a proposal before commitment.
The school should not present only two options: accept the entire tool or have no solution. It could compare manual registration, cards, different procedures and versions collecting less data. This comparison makes discussion more concrete. A feature that is optional for a supplier can create an obligation for a student. The relevant boundary is found not only in a feature list but also in the relationships that adoption creates between the people involved.
Participation must have consequences
Inviting people to a meeting does not guarantee influence over the result. A useful consultation provides understandable information, explains what can still change and records how objections were considered. If the decision is already final, calling its presentation participation can create a false expectation. Listening also requires attention to people who lack the time, access or technical vocabulary to follow the process. An unanswered concern should remain visible rather than disappear from the record.
Clear rules and opportunities for revision
A responsible decision can specify where use is allowed, which functions are excluded, who monitors results and which events require suspension or review. This avoids relying entirely on operators' promises to exercise good judgment. An experimental period needs a defined duration, assessment criteria and an exit. Experimentation without stopping conditions can turn a temporary arrangement into permanent adoption through inertia. A review date is useful only when someone has both responsibility and authority to act on the findings.
Who answers when something goes wrong?
Before deployment, describe a plausible error and follow its path: who notices, who receives the complaint, who can correct it and who explains the decision? When every participant points elsewhere, responsibility disappears in practice. Contracts and procedures need to enable action, not merely attach names to a chain. An affected person should be able to obtain an understandable response without mastering the service's technical architecture or negotiating separately with every supplier.
Frequently asked questions
Should specialists decide everything?
No. Their assessment is essential for technical questions, but decisions about objectives and risk distribution also need legitimacy and participation.
Do limits prevent innovation?
Not necessarily. Clear limits can guide experiments, expose responsibilities and allow justified changes when new evidence emerges.
