Marketing needs
August arrives and the pace changes.Part of the team is on holiday, meetings become less frequent,…
We increasingly make decisions with the help of an automatic recommendation.
A platform chooses which video we will watch next. An app suggests how to reply to a message. An assistant drafts the email we need to send. A shop highlights the product that, in theory, suits us best. Even when we look for a route, a news story or a restaurant, we rarely start from a blank page: the system has already arranged the options and placed one of them before us.
In many cases, this help is useful. It saves time, simplifies processes and prevents us from spending energy on unimportant decisions. It would make no sense to reject a tool simply because it makes a choice easier for us.
However, this convenience can also conceal a deeper shift. Every time we accept a recommendation without examining it, we stop making a small decision. A single delegated choice may seem insignificant, but the habit of systematically accepting the suggested option can gradually change our relationship with judgement, doubt and responsibility.
The boundary is not always obvious. We may think that we are still deciding because we are the ones pressing the button, confirming the reply or making the purchase. But choosing between options that have already been filtered, ordered and presented by a system is not quite the same as building a decision from the start.
The question, then, is not whether these tools are useful. They are. The issue is at what point a recommendation stops helping us decide and starts deciding for us.
Automating a task is not delegating a decision
Automating a task and delegating a decision may seem like similar actions, but they do not mean the same thing.
Automating means entrusting a tool with carrying out something we have already decided. We know what we want to achieve and use technology to do it more quickly, accurately or efficiently. We can ask an app to calculate the shortest route, a spreadsheet to sort some data or an assistant to summarise several documents. In all these cases, the tool makes the process easier, but the objective remains ours.
Delegating a decision means going one step further. We no longer merely ask for an instruction to be carried out, but for the system to determine what should be done. It does not just calculate how to get somewhere, but where we should go. It does not just compare several options, but points out which one we should choose. It does not just draft a reply, but defines what position we should take.
The difference lies in who sets the criteria.
When we automate, the tool works within a prior decision. When we delegate, we allow it to take part in defining the decision itself. And although that boundary is not always visible, it is important to recognise it.
A recommendation can help us assess alternatives, identify information we had overlooked or reduce the complexity of a choice. The problem arises when we stop using it as support and begin treating it as a definitive answer.
Because a tool can calculate which option is faster, cheaper or more likely. What it cannot determine on its own is what is worthwhile, what best aligns with our values or what responsibility we are willing to take on.
Systems that learn what we are likely to choose
Automatic recommendations do not appear neutrally. Behind every suggested piece of content, highlighted product or proposed reply is a system trying to anticipate which option is most likely to interest us.
To do so, it analyses our behaviour: what we have watched, how long we have spent on a page, what we have bought, what we have ignored, what we have searched for or what decisions we have made before. It also compares this data with that of other people who show similar patterns and uses that information to order the options it presents to us.
But the system does not only try to understand us. It also responds to the objectives of the platform that designed it.
A recommendation may be optimised to keep us connected for longer, get us to buy a product, click on a link or consume more content. What appears first does not necessarily have to be what is best for us, but rather what the system considers most effective at producing a particular response.
The more precise these recommendations become, the less visible their influence is. We do not feel that someone is deciding for us. We still have the possibility of rejecting the suggestion, looking for another option or leaving the platform.
However, the recommended option already has an important advantage: it appears first, seems more relevant and requires less effort than the others.
Systems do not need to force us to choose. It is enough for them to place one possibility in front of us at the right moment. And when an option fits well enough with what we expected to find, the likelihood that we will continue searching decreases.
In this way, influence does not operate through obvious imposition, but through selection, order and timing. The decision still appears to be ours, even though the space in which we decide has been organised beforehand according to someone else’s criteria.
The risk of always accepting the first option
The first option often has an advantage that does not necessarily depend on its quality: it requires the least effort.
The first generated text may seem correct enough. The first suggested reply may fit what we wanted to say. The first recommended product may meet our requirements. The first plausible explanation may resolve a doubt. The route an app presents as optimal may save us a few minutes.
On many occasions, accepting that suggestion is reasonable. Not every decision requires exhaustive research or a comparison between multiple alternatives. Choosing the first option is not, in itself, a problem.
The risk arises when it stops being an occasional choice and becomes a habit.
When we systematically accept the first suggestion, we may lose the habit of reviewing, checking and formulating alternatives. We stop asking ourselves whether the text truly expresses what we want to say, whether the reply represents our position, whether the recommended product is the one that suits us best or whether the explanation we have received considers other possibilities.
The first option then works as a default answer. We do not choose it because we have concluded that it is the best, but because it is already available, seems suitable and saves us the effort of thinking further.
This mechanism can be especially difficult to spot when the suggestion is good. A clearly mistaken recommendation prompts doubts. A reasonable option, by contrast, invites us to accept it without further thought. And precisely for that reason, it can have more influence: it does not need to convince us completely, only to seem valid enough.
Retaining the ability to decide does not mean distrusting every suggestion or automatically rejecting the first alternative. It means keeping alive the habit of asking why that option should be accepted, what criteria support it and what we might be leaving out by not considering other possibilities.
Dependence that does not look like dependence
When we think of technological dependence, we tend to imagine an obvious incapacity: not knowing how to find our way without an app, being unable to write without an assistant or being incapable of completing a task without turning to a tool.
But dependence does not always appear when we lose a capability. It can also begin when we still retain it, but stop using it.
We may still be capable of comparing options, drafting a reply, seeking information or making a decision on our own. However, doing so requires more time, more attention and, often, more discomfort. Faced with an immediate recommendation, thinking for ourselves can seem unnecessarily slow.
That is where convenience begins to transform our behaviour.
Every delegated decision seems insignificant. Accepting a suggested reply, following a recommendation or choosing an already prepared option hardly changes anything on its own. But when that gesture is repeated, we can become accustomed to receiving almost finished decisions.
Judgement also needs exercise. It grows stronger when we compare, doubt, interpret and take on the consequences of choosing. If we stop practising these actions because a tool resolves them first, our ability to decide does not disappear overnight, but it can become less active, less demanding and more dependent on external prompts.
The quietest form of dependence is not being unable to decide. It is no longer feeling that it is worth doing so.
And when living among ready-made options becomes normal, convenience stops being merely an advantage. It begins to define how we relate to our own freedom.
A recommendation does not remove our responsibility
Just because a tool proposes an action does not mean it assumes its consequences.
A system can recommend whom to hire, which patient needs more attention, which student needs support, which message should be sent or which priority should come first. It can organise information, detect patterns and suggest a response. But the final decision still has effects on real people.
That is why responsibility does not disappear when an algorithm is involved.
We may feel that a choice is more objective because it has been calculated by a tool, but every system works from data, criteria and objectives defined in advance. It can also reproduce errors, overlook nuances or fail to understand circumstances that are essential to interpreting a situation correctly.
When a decision affects other people, it is not enough to know which option the system recommends. We also need to ask why it recommends it, what information it has taken into account, what elements it has left out and who will bear the consequences if it gets it wrong.
This is especially important in decisions involving values, relationships or potential harm. Hiring or dismissing someone, assessing their performance, educating a child, making a diagnosis, caring for someone vulnerable or communicating sensitive news are not simple optimisation problems.
In these cases, deciding requires interpreting the context, listening, recognising exceptions and considering aspects that cannot always be turned into data. It also requires empathy, prudence and the ability to answer to the person affected.
Artificial intelligence can provide valuable information and help us identify elements we had overlooked. However, it should serve to broaden our deliberation, not replace it.
The more important a decision is and the greater its consequences, the less reasonable it is to accept a recommendation without understanding it. We can rely on a tool, but we should not use it as an excuse to avoid the responsibility of choosing.
Conclusion
Using recommendations does not mean automatically giving up our freedom. Tools can help us organise information, reduce complexity and find options we might not have considered on our own.
The problem is not receiving help, but ceasing to take an active part in the decision.
Retaining the ability to choose requires maintaining some simple habits: asking for alternatives, reviewing the criteria used, understanding what objective the system is optimising and pausing when a decision has important consequences. It also means being able to explain why we have chosen one option rather than another.
Not all decisions require the same level of reflection. We can accept a suggested route, a content recommendation or an automatic reply without turning every everyday action into a complex exercise. But the more a decision affects our values, other people or the direction we want to take, the less sense it makes to delegate it without scrutiny.
The issue is not making every decision without help. It is preventing the ease of receiving an answer from replacing our responsibility to decide.
Because deciding is not only about choosing between options. It also means recognising what we are prioritising, what consequences we are willing to accept and what kind of person, organisation or society we want to become.
We can delegate calculations, searches and processes, but we should not delegate just as readily the responsibility of deciding what is worth doing.
What decisions are you delegating for convenience that you should still be able to explain as your own?
Sources:
Comments