A Computer Science Researcher At Aston University Has Used Artificial Intelligence (AI) To Show That We Are Not As Individual As We May Like To Think

The influence of one’s peers significantly affects individual actions. The dynamics of a group affect its members’ propensity to break the law, use violence, or aid those in need.

Studies have shown that looking at a group of people has a powerful effect on people’s focus. The things we pay attention to significantly impact how we react. The conventional explanation is that this behavior is adaptive; when we observe many people fixating on the same object, we reason that this must be significant and we decide to follow the group’s gaze.

Study after study on gaze cueing has shown that people are quicker to detect stimuli when they are preceded by a face looking in the direction of the stimulus. This holds true even when the gaze direction is unpredictably opposite the stimulus’s location. In a broader sense, studies on automatic imitation have revealed that people often unconsciously mimic harmful behaviors.

However, research on gaze cueing and automatic imitation is typically conducted in a controlled laboratory setting using fabricated computer tasks that indirectly measure obedience to social norms. Therefore, it is challenging to extrapolate from these studies the influence of social groups in the real world. So it is unclear if strategic processes are the only ones involved in social group impact or if reflexive processes also play a role. 

New research by Aston University uses VR to answer this question by improving ecological validity while maintaining experimental control. In particular, the group of researchers designed an interactive VR task in the style of gaze cueing and automatic imitation experiments. 

In their experiment, ten virtual agents joined participants while they watched a video outside. Participants were given the option of either being asked to look up to the left (a forced choice) or the right (a forced choice) or having their gaze split evenly between the two options (a 50/50 split) (free choice). In addition, anything from one to ten different agents simultaneously cast their gaze skyward, either to the left or right. Participants were required to find a fire that they were told would be present in both settings to finish the experiment. In this way, they could isolate the impact of reflexive, bottom-up processes from indirect and direct measures of gaze following, given the gaze direction of the virtual agents lacked any information about the target position.

If strategic processes aren’t the only ones contributing to social group influence, then the number of virtual agents looking up, even when the group has no relevant information, should increase. The team used forced choice trials to develop an indirect measure of gaze following. It classified congruent trials, in which participants were required to glance in the same direction as the virtual agents, eliciting faster responses than incongruent trials. On the other hand, the direct measurement of gaze following was derived from free-choice trials. It was understood as the percentage of trials in which participants elected to follow the agents. 

The experiment’s findings corroborate the idea that a two-stage model best describes the effects of a large group of people. This means that people are naturally inclined to imitate the behavior of those around them. A more deliberate, strategic process follows when an individual decides whether or not to emulate those in their immediate environment. This impact is felt not just through social norms but also immediate actions and is at the core of collective behaviors like rioting and mass panic.

This Article is written as a research summary article by Marktechpost Staff based on the research paper 'Evidence for a two-step model of social group influence'. All Credit For This Research Goes To Researchers on This Project. Check out the paper and reference article.

Please Don't Forget To Join Our ML Subreddit