Agency: when understanding is no longer enough

Why this matters now

Plausible answers, alternative options, images, drafts and preliminary analyses can increasingly be generated at very low friction. That abundance does not remove human responsibility; it changes where responsibility concentrates. If a system can produce ten possible directions in seconds, someone still has to decide which direction deserves to become real, what evidence is sufficient, which trade-offs are acceptable, who may be affected and what happens when the first decision fails.

The Real-Life Readiness framework places this transition at the centre of human preparedness. Knowledge and access to answers are necessary but not sufficient. Readiness becomes visible when a person can mobilise what they know, connect it with what others know, face uncertainty, decide, act and take responsibility for consequences. Agency matters because it prevents the learner, worker or citizen from becoming only a receiver of instructions, algorithmic outputs or decisions made elsewhere.

Mars Challenge turns that position into an educational design principle. The point is not only to expose young people to complex problems or to reward an attractive idea. The process creates moments in which a team must choose and defend a direction. Learning is no longer neutral at that point: understanding creates a demand for judgement. Judgement becomes Agency when there is enough authorship to choose and enough responsibility to answer for the choice.

What Agency means

Agency names a human condition: the capacity to intervene deliberately in a situation rather than merely receive it. Mars Challenge materials use a concise formulation — authorship under real consequences. The phrase matters because it binds influence and responsibility together. A person exercises Agency when they materially participate in determining what happens next and cannot separate that participation from accountability for what follows.

Four elements belong together. The first is understanding. Action without understanding may be impulse, compliance or improvisation; it is not competent Agency. The second is judgement: evaluating evidence, alternatives, values and limits when information is incomplete. The third is choice and action: something must genuinely be selected, prioritised, built, rejected, defended or changed. The fourth is responsibility: the choice extends beyond the moment of decision into its effects, revision and accountability to people and systems affected by it.

Agency is therefore not the same as motivation. A highly motivated person can still be executing someone else’s decisions. It is not absolute autonomy, because responsible choices always exist within relationships, rules and constraints. It is not confidence: a person can exercise Agency while openly recognising uncertainty. And it is not identical to formal leadership. Someone without a leadership title can exercise Agency by identifying a risk, challenging an assumption, introducing missing expertise or causing a justified change of direction.

Within Mars Challenge, Agency also describes a shift from being evaluated to becoming an author. When knowledge is mainly something a learner reproduces for someone else to judge, authorship is limited. When participants must decide which problem matters, what information is reliable enough, which solution should advance and how to respond to consequences, knowledge becomes material for judgement and action.

From understanding to responsible authorship

1. Understanding is not yet deciding

A person can understand a problem deeply and still change nothing. They may explain an environmental, social or technological risk in detail without ever confronting a meaningful choice. Agency begins when understanding reaches a fork: more than one path is possible, and no path can be selected by simply reproducing a memorised answer.

Many learning environments end before this threshold. They ask students to identify, explain, describe or analyse. Those capacities matter, but they do not necessarily require the learner to take a position. Mars Challenge adds another condition. Teams choose a challenge, converge on an idea, allocate responsibilities, define a solution, build a prototype, respond to feedback and justify an Earth return. These actions do not prove that Agency has increased as a psychological trait, but they create situations in which authorship can be observed.

If there is no meaningful choice, there is limited space in which Agency can appear. A fully prescribed activity may teach valuable knowledge, technique or discipline, but most authorship remains with the person who designed the instructions. To make Agency visible, a consequential part of the route needs to remain genuinely open to participants.

2. Judgement comes before responsible choice

Agency does not make every participant decision a good decision simply because the participant made it. Confidently choosing with weak evidence is still weak judgement. The movement from understanding to action therefore needs an evaluative layer.

In Real-Life Readiness, DECIDE means exercising judgement with incomplete information, competing values and trade-offs, and determining what is reliable enough to act on. This avoids two opposite mistakes: waiting for impossible certainty and mistaking speed for competence. Real problems often do not provide complete information on demand, but uncertainty does not justify acting without criteria.

Competent Agency asks: What do we actually know? What are we assuming? What requires verification? Whose expertise is missing? Which criterion has priority? Which cost are we accepting? Which risk are we not entitled to externalise? These questions turn preference into explainable judgement.

Generative AI makes this layer more, not less, visible. A system can produce alternatives, arguments and artefacts, but it does not automatically carry responsibility for the context in which they will be used. The Real-Life Readiness position is that human work increasingly concentrates on problem framing, verification, contextualisation, choice and consequence. From this perspective, “the AI decided” is not an adequate account when a person or institution still authorises implementation.

3. Choice includes giving something up

A real choice closes other possibilities. Selecting one idea means not advancing another. Prioritising one need means allocating less attention elsewhere. Finite resources force teams to decide where time, materials, energy and expertise will go. Constraints can therefore make Agency more visible because they remove the option of keeping every path open indefinitely.

Mars Challenge uses constraints as part of its mission environment: limited resources, extreme conditions, challenge requirements, team roles and deliverables. The educational value is not difficulty for its own sake. It is the appearance of trade-offs. When a team cannot do everything, it must show what it considers essential and why.

This is where choice becomes responsibility. Decisions reorganise shared resources and affect other people’s work. The more a decision changes the common system, the more important it becomes to explain and own it.

4. Action turns a position into an intervention

An intention remains relatively protected until it meets reality. ACT, in Real-Life Readiness, means turning knowledge into a choice, intervention, prototype or commitment and adapting it through feedback. This transition matters because reality generates information that isolated reasoning cannot provide.

A prototype may fail. A user may interpret it differently from what the team expected. A new fact may undermine an assumption. A technically elegant solution may fail because of cost, adoption, context or resource constraints. Agency is not defending the first decision at all costs. It is retaining enough authorship to revise action when the evidence changes.

The Mars Challenge process makes this relationship concrete through prototyping and feedback. Teams do not only state what should work; they represent a solution, show it, receive response and adjust. Their reasoning leaves traces: versions, decisions, justifications and revisions. These artefacts make it possible to examine how judgement became action.

5. Ownership continues after the decision

Agency would be incomplete if it ended at the moment of choice. Mars Challenge explicitly connects it with consequences, and Real-Life Readiness adds CARE: taking responsibility for effects on people, relationships and living systems on which action depends.

Ownership does not mean controlling every consequence. Complex systems make that impossible. It means that lack of total control does not remove the obligation to anticipate, observe, respond and correct. An agentic response can say: “This was our decision, these were our criteria, this effect was not anticipated, and now we need to revise.”

Responsibility is therefore different from omniscience. It is continuity. It rejects the pattern in which people claim authorship of success while treating harmful consequences as belonging to somebody else.

6. Learning closes the loop without ending it

Consequences return information to the decision-maker. After action, the question is no longer only “Were we right?” It becomes “What did reality show us?”, “What changed?”, “What would we revise?” and “What responsibility is still open?”.

Agency is not a heroic one-off moment. It is a recurrent practice. The same person may show strong Agency in one context and need support in another. They may have deep technical knowledge but miss social consequences. They may act decisively with insufficient evidence. They may take responsibility and still need a different form of expertise.

For explanatory purposes, this article uses a six-movement editorial synthesis: understand → judge → choose → act → own → learn. It is not an official replacement for the 12 Mars Challenge phases, nor is it presented as a validated psychometric scale.

Agency is not autonomy, motivation, confidence or leadership

Autonomy concerns the space a person has to self-direct. Agency needs meaningful autonomy, but it adds judgement and responsibility. Choosing a cosmetic detail can involve autonomy without consequential authorship. Choosing which problem to pursue and answering for the decision can involve Agency.

Motivation concerns energy or willingness to act. High motivation can exist under complete prescription. A person can enthusiastically execute someone else’s plan. Conversely, Agency may be required when motivation is low but responsibility cannot be delegated.

Confidence is a belief about one’s capacity. Agency does not require certainty. In complex problems, saying “we do not know yet; this needs verification” can demonstrate stronger judgement than moving forward with unjustified confidence.

Leadership often requires Agency, but Agency is not reserved for leaders. Mars Challenge distributes roles across strategy, user understanding, communication, technical viability, prototyping and planetary coherence. Any of those roles can change the team’s direction by introducing evidence or exposing a weak assumption.

Activism can be an expression of Agency, but the concept is broader. Agency may also mean revising a scientific hypothesis, rejecting an unviable solution, pausing implementation, acknowledging a team error or adapting an intervention after new evidence.

The common denominator is not emotion or title. It is the relationship between judgement, choice, action and responsibility.

What the evidence says

The available evidence base combines conceptual and operational sources, and the distinction matters.

First-party conceptual source. THE AGENCY 2026 formalises Agency within the Mars Challenge ecosystem as a human state. It places the threshold at the point where understanding limits is no longer enough and understanding must translate into deliberate capacity to redesign and act responsibly. The Mars Challenge 2026 framework/manifesto reinforces this by describing Agency as authorship under real consequences rather than motivation alone.

Conceptual framework. Real-Life Readiness provides a broader architecture: KNOW, CONNECT, FACE, DECIDE, ACT and CARE. It describes Agency as beginning when a person moves from being primarily an object of evaluation to using knowledge to shape a path and contribute to change. It also identifies authorship, constraint, consequence, reflection and transfer among the conditions that can make preparedness visible.

Documented implementation. The 2026 Participant Guide records actual decision points in the designed process: recognising strengths, taking roles, choosing an idea, explaining convergence, defining problems, prototyping, receiving feedback, revising a solution and articulating Earth impact. This demonstrates that the programme design contains opportunities for meaningful choice and responsibility inside a bounded learning environment.

The boundary of the evidence is equally important. The supplied corpus does not contain a validated Agency scale, control group, pre/post measures or longitudinal study. It would therefore be incorrect to claim that Mars Challenge increases Agency by a quantified amount or that Agency has been independently validated as a standalone causal construct. The appropriate evidence status is Partial: the definition and operational conditions are documented; causal effects remain a research question.

What this means for young people, educators and institutions

For young people, Agency changes what success looks like. The task is no longer only to reach an impressive answer. It includes being able to explain what decision was made, why, what evidence supported it, what was rejected, who may be affected and what will happen if the assumption fails. Knowledge becomes more important, not less, because judgement without knowledge is shallow.

For educators and mentors, designing for Agency does not mean removing structure. Unbounded freedom can simply privilege participants who already know how to navigate the context. The challenge is to create meaningful autonomy inside clear boundaries: authentic problems, real choices, transparent criteria, finite resources, feedback and revision. The educator becomes an architect of judgement conditions as well as a source of knowledge.

For institutions, Agency changes what counts as evidence of learning. A polished final prototype reveals little about who made decisions, which alternatives were considered, what disagreement changed the route or how the team responded to failure. If Agency matters, institutions need process traces: decisions, rationales, revisions, responsibility allocation and considered consequences.

In AI-rich environments, the implication is direct. Delegating generation cannot mean delegating responsibility. Institutions can permit generative systems while still requiring participants to identify what was generated, what they verified, what they rejected, what context they added and why they remain accountable for the final decision.

Related concepts

• Real-Life Readiness — the broader category in which judgement, action and responsibility operate as an integrated human capacity.
• Earth-Back Thinking — introduces constraints and systemic consequences that make decisions less abstract.
• Next Human Learning — frames judgement, adaptation and responsibility as critical forms of human learning in an intelligence-rich environment.
• Authorship — the capacity to influence the path materially rather than simply execute a hidden correct answer.
• Responsibility — the continuity of a decision into its effects, revision and accountability.

Evidence in practice

No individual project is labelled here as causal proof of Agency. The supplied documentation supports operational proof of design, not a validated psychological outcome for a specific team.

1. Agency definition in THE AGENCY 2026 and the Mars Challenge 2026 framework/manifesto.
2. DECIDE → ACT → CARE and authorship/consequence conditions in Real-Life Readiness.
3. Documented sequence of decisions, responsibilities, prototyping, feedback and revision in the 2026 Participant Guide.

Frequently asked questions

Does Agency simply mean taking initiative?

No. Initiative can be part of Agency, but moving first does not demonstrate sound judgement. Agency connects action with understanding, criteria and responsibility. The relevant questions include why the action was taken, which alternatives were considered, what evidence supported it, which consequences were anticipated and how the person or team will respond. Initiative without judgement may be impulse; initiative with responsible authorship can become Agency.

What is the difference between Agency and autonomy?

Autonomy describes the space available to choose. Agency describes the responsible exercise of that space when the choice matters. A learner may have freedom to choose superficial details while having no influence over the real direction. Agency requires materially meaningful choice and a relationship with consequences. It does not require unlimited freedom: it can be exercised within rules, constraints, roles and shared goals.

Can Agency be taught directly?

The supplied evidence does not demonstrate that Agency can be taught as an isolated content unit, nor does it establish a validated causal relationship. The Mars Challenge and Real-Life Readiness frameworks propose that environments can be designed in which Agency must be exercised: authentic problems, open decisions, constraints, consequences, feedback and reflection. Whether those conditions produce durable, measurable change remains an empirical research question.

How are Agency and Real-Life Readiness related?

Real-Life Readiness is a broader category of human preparedness. Agency runs especially through DECIDE, ACT and CARE, while depending on KNOW, CONNECT and FACE. A person needs to recognise what they know, connect with other knowledge and face uncertainty before deciding responsibly. Agency describes the movement from that knowledge into authorship and action, but it does not exhaust the full Real-Life Readiness framework.

Why does AI make Agency more important?

Because generating an answer and deciding what should become real are different tasks. An AI system can provide plausible options, but the user still needs to verify, contextualise, establish criteria, weigh trade-offs and own consequences. In this framework, AI can expand the option space while Agency governs how that space is converted into responsible action.

How can Agency be observed without reducing it to a test?

Real-Life Readiness proposes behavioural and artefact-based evidence: how a person frames a problem, distinguishes evidence from assumption, identifies missing knowledge, explains trade-offs, distributes responsibility, revises after feedback and responds to consequences. These traces can document the process. They should not automatically be treated as a validated psychometric measure of Agency.

References

THE AGENCY 2026 Mars Challenge / Virtual Educa / Space Talent Lab 2026 First-party document supplied in the Mars Challenge corpus.

Mars Challenge 2026 (tierra) new — Program Framework / Manifesto Mars Challenge 2026 First-party document supplied in the Mars Challenge corpus.

Real-Life Readiness: A Foundational Framework Adelino Sousa / Mars Challenge / Virtual Educa 2026 Foundational document supplied in the corpus; defines KNOW, CONNECT, FACE, DECIDE, ACT and CARE and explicitly states its non-psychometric conceptual status.

Guía del Participante · Mars Challenge · The Grand Jam 2026 Mars Challenge 2026 First-party operational guide documenting roles, decisions, prototyping, feedback and Earth return.

Explore next

1. Real-Life Readiness — explore the broader category of human preparedness in which Agency operates.
2. Earth-Back Thinking — see how extreme constraints make choices, trade-offs and consequences visible.
3. Mars Challenge Methodology — understand where authorship, decision, prototyping, feedback and responsibility enter the process.

Next step

Deepen: Explore Real-Life Readiness

See evidence: See how decisions appear in the process and projects

Participate: Discover how to take part