The Art of Deciding When the Answer Isn’t Clear

When certainty is impossible, what makes a good decision?

Complete certainty rarely arrives before an important decision has to be made. Waiting may feel responsible, but inaction can shape the outcome just as decisively as action. The challenge is to combine credible evidence with experience and judgement, while remaining alert to bias and ready to change course.

This was the central theme of Who Decides: The Science of Decisions, the 14th anniversary seminar of the Estonian Research Council’s Brussels Office, held on 22 September at the Permanent Representation of Estonia to the European Union. The discussions moved from everyday habits, confirmation bias and AI to high-level policymaking, where evidence is often incomplete, time is limited and responsibility cannot be delegated.

In her opening remarks Estonian Research Council Director General Dr Kristiina Tõnnisson compared the office’s 14th anniversary to adolescence, a stage when authority begins to be questioned. The metaphor captured a central question of the seminar: who gets to decide, on what basis, and when should established assumptions be challenged?

Five ideas worth taking from the seminar

Rationality depends on context

The discussion challenged the simple division between rationality and irrationality in decision-makers. How carefully we think depends on urgency, habit, attention and the task itself. Fast, automatic thinking is not inherently flawed, and intuition may reflect years of accumulated expertise. The useful question is what lies behind a gut feeling and whether experience gained in one context is relevant in another.

“What we perceive as gut feeling is often some mixture of preference, intuition or expertise,” said Dr Silja-Riin Voolma, applied behavioural scientist.

Veronika Kalmus, Professor of Sociology, added that the information environment shapes what appears relevant, trustworthy and rational. Media agenda setting affects which issues and voices people encounter, while confirmation bias makes us more receptive to information that supports what we already believe.

AI relocates the human problem rather than removing it

“Humans can quite quickly fall into being the rubber stamp that simply puts their name under it,” said Uku Sildoja, Early Stage Researcher at Next Gen Digital State Research Group at Tallinn University of Technology.

AI can process information at a remarkable scale, but it does not eliminate bias, uncertainty or responsibility. Sildoja compared black-box AI with human gut feeling: both may produce a useful answer without making the reasoning fully visible.

Sildoja argued that adding a human reviewer is not, by itself, a sufficient safeguard. Reviewers need the time, expertise and practical ability to question or overrule a system’s output. He advocated designing systems that can automate suitable routine cases, while reserving expert attention for difficult or consequential decisions and providing meaningful routes to challenge and escalation. Comparing AI to a new employee, he suggested granting greater autonomy as reliable performance is demonstrated, alongside continued monitoring and spot checks.

Evidence must enter policymaking early

Research has the greatest influence when it helps define the problem, identify options and establish the criteria for success. If experts are consulted only after a political position has hardened, evidence risks becoming justification for a decision already made.

Maive Rute, Deputy Director-General of the European Commission’s Joint Research Centre, described the Joint Research Centre as one institutional response to this challenge. Its purpose is to bring together internal and external expertise, test models and synthesise large bodies of knowledge into forms that policymakers can actually use.

“Research should not be like a spice put on food, but a natural part of policymaking,” said Maive Rute.

Good evidence alone does not guarantee influence. Whoever sets the agenda, writes the first draft, advances an argument or builds a coalition can shape the outcome. Evidence-based policymaking therefore requires institutions that give knowledge a meaningful route into the decision-making process.

Decision-makers need usable expertise, not only excellent science

Scientists are expected to qualify their conclusions and identify what remains unknown. Political leaders, by contrast, must choose a course of action at a particular moment. Science can narrow uncertainty and test assumptions, but it rarely produces one unavoidable policy choice.

“An excellent scientist is not necessarily a good expert for decision-makers,” saidJüri Luik, former Estonian Minister of Foreign Affairs and Minister of Defence.

The challenge is therefore to translate complexity without erasing it. A useful adviser must be able to identify the relevant facts, explain the principal uncertainties, outline realistic options and clarify their consequences. This is particularly important in foreign and security policy, where events can move faster than research and where the independence of expertise must itself be examined.

Not deciding is also a decision

In high-stakes situations, waiting for uncertainty to disappear is rarely realistic. Delay may close options or allow others to determine the outcome. Decision-makers therefore need to judge when the available evidence is sufficient to act, state the assumptions behind the decision and consider what new information would justify changing course.

As Matti Maasikas, Deputy Secretary General at the European External Action Service, noted, the more a position is grounded in previous research and evidence, the stronger the basis for action is, but decisions must ultimately be made within real-world deadlines.

This matters especially in large institutions, where reversing a collective decision may be harder than making it. Adaptability should be built into the process from the start.

“You can decide not to do anything, but then it has to be a deliberate decision,”added Jüri Luik.

The central lesson

Good decision-making is not a choice between science, experience and judgement. It requires all three, along with a clear understanding of their limits. The goal is not the mythical perfect decision, but disciplined judgement: knowing what is known, recognising what remains uncertain, acting when the available evidence is sufficient and preserving the ability to change course.

Explore highlights from the event and view the full photo gallery on Flickr.