The example taken here is that of a complex model for managing weight loss, containing both relevant and irrelevant information. When relevant information is not highlighted in the model, participants make poor decisions. More broadly, when we have to make a difficult decision, our first instinct is usually to gather as much information as possible. However, with too much information, decision-making deteriorates rather than improves, and increasingly so with additional detail.
One says “make informed decisions”
But in fact, it would rather be a matter of taking them on the basis of a minimum of relevant information.
“Too much information kills the decision”,
“this principle is counterintuitive, because we all like to think that we use information wisely to make intelligent decisions”says lead author Samantha Kleinberg, associate professor of computer science at Stevens Institute of Technology. “But the reality is that when it comes to information, more is not necessarily better. »
The study which deciphered how most people make decisions, reveals that most focus on the information provided to them. It doesn’t matter whether they are relevant or not. This work shows in particular that when it comes to everyday scenarios, such as making healthy decisions about nutrition, people’s ability to reason effectively evaporates quickly. “’Superfluous’ knowledge and beliefs distract people from the rational model: when I think about what I am going to eat, for example, I can have all kinds of preconceived ideas about which foods to favor and ultimately this disrupts the data to actually take into account in this decision making, however simple and everyday.
A series of experiments exploring how decision-making changes not only depending on the object, ranging from buying a house to weight management, the causal model, but above all depending on the information provided:
- even a tiny amount of irrelevant information has a significant negative effect on decision making;
- the more excess or irrelevant information there is, the less relevant the decision;
- with too much information, decision-making quickly becomes as bad as if no information had been provided.
Ex : if a causal model shows that consumption of salty foods increases blood pressure, but also provides extraneous information, such as the fact that drinking water decreases thirst, it becomes much more difficult to make the right choice about consumption salty foods.
In real lifethe researchers add, not only are people overwhelmed with information, but they have difficulty determining “which part of the model” they should pay attention to – in other words, what is the objective being pursued.
These observations have implications for public health educational messages which should only target a simple objective and indicate the essential measures to achieve it.
“The public needs simple, carefully targeted causal models to make good health decisions.”
In this regard, AI-based chatbots open up a new opportunity to deliver just the right health information based on individual characteristics and goals.