How avatars are making people take more risks

In a world shaped by virtual interactions, avatars are now the new face of communication. In games, work, shopping, or therapy, they act like real people – showing expressions and emotions, and even changing how we feel about taking risks.

But what happens when our brains start treating these digital stand-ins differently than real humans?

A study by Toshiko Tanaka and Masahiko Haruno of the National Institute of Information and Communications Technology (NICT) in Japan has brought this question into sharp focus.

The research explores how avatars influence human behavior, particularly when it comes to taking risks, and it offers a deeper look at the neural mechanics behind this shift.

More risks with avatar feedback

Tanaka and Haruno designed a study that places human decision-making under the microscope in virtual environments.

Participants engaged in a gambling task, where they had to choose between a guaranteed small reward and a risky option with a larger potential payout. Their choices were followed by dynamic facial feedback from either a real human face or an avatar presenting the same expressions.

What the researchers found was striking: people made more risky choices when they expected to receive feedback from an avatar rather than a real face.

The difference was not subtle. It persisted across both behavioral and neuroimaging experiments, revealing that avatar interactions fundamentally altered the way people assessed the value of risk.

The brain’s reaction to avatars

The answer lies in how our brain interprets uncertainty, especially in social situations. The researchers discovered that avatars shift our emotional processing of risk.

Participants perceived facial feedback from avatars – be it admiring or contemptuous – as less emotionally charged than the same expressions from real humans. This led to a more favorable valuation of feedback uncertainty.

That means when participants weren’t quite sure what kind of response they’d receive, the ambiguity felt less threatening if it came from an avatar. This change in emotional processing allowed them to focus more on potential gains, encouraging riskier choices.

Quieting the brain’s emotional alarm

To understand the brain’s role in this behavior, the researchers used functional magnetic resonance imaging (fMRI) scans.

The experts found that a key emotional processing center in the brain – the amygdala – was deeply involved. Specifically, activity in the central amygdala (CM) decreased when feedback uncertainty was delivered through avatars.

This reduction in amygdala response reflected a lower emotional cost attached to avatar feedback. In essence, avatars made the stakes feel lower, even when the actual risk remained the same.

This neural response was not random. It tracked closely with how individuals evaluated feedback uncertainty and how often they chose the risky option.

Empathetic people are more affected

Not everyone responded the same way to avatar feedback. The researchers included psychological assessments that measured traits like empathy and anxiety.

They found that people who scored high on empathetic concern – those more emotionally attuned to others – were especially likely to show reduced amygdala responses and increased risk-taking in the avatar condition.

The central amygdala response to feedback uncertainty also correlated specifically with these empathy scores. It wasn’t general anxiety or social avoidance that shaped the brain’s behavior – it was the emotional consideration of others’ internal states.

According to the study, this pattern was not observed in participants with higher scores for social anxiety. That distinction suggests that avatars reduce emotional burden without diminishing the social component of the interaction.

Other brain areas that responded

Though the amygdala played the primary role, other brain regions also responded. The ventral striatum (vSTR) and ventral anterior cingulate cortex (vACC) showed activity linked to feedback uncertainty, though their involvement was less direct.

These areas processed reward-related and emotional information, but they didn’t show the same predictive power as the amygdala when it came to actual behavior changes.

Participants who took more risks with avatars had amygdala activity patterns that clearly differed from those who remained cautious. The vSTR and vACC supported these shifts but did not appear to drive them.

How avatar influence was measured

The strength of these conclusions rests in a meticulous experimental design. Participants interacted with a same-gender observer whom they believed was monitoring and evaluating their gambling behavior.

Unbeknownst to them, feedback expressions were pre-recorded, and observer appearances alternated between avatar and real-face conditions.

Each trial in the experiment involved choosing between a safe and risky option, followed by feedback in the form of facial expressions. The researchers recorded behavioral data and brain activity, allowing them to model the valuation process mathematically and neurologically.

Computational modeling confirmed that feedback uncertainty carried more weight in decision-making under avatar conditions. The best-fitting models showed that only this parameter needed to vary between avatar and human feedback to explain participants’ risk behavior.

Tests confirm risk-taking boost

To ensure their findings were not due to a simple lack of engagement, the team ran pilot tests. These included a no-observer condition, in which participants received no feedback at all.

Risk-taking rates in this condition were significantly lower than those seen with avatar feedback. This helped eliminate the possibility that participants were simply ignoring avatar responses.

Even when compared to human feedback, avatars elicited more risk-taking. This demonstrated that it wasn’t just the presence of feedback but the form of the feedback that mattered.

Real-world implications

“We found that people tend to take more risks when their partner responds through an avatar rather than showing their real face. This seems to be driven by a change in how they process uncertainty – and interestingly, that change is reflected in the amygdala,” said Haruno.

“Using the same confederate and having them pass as just another participant every time wasn’t easy – it took a lot of effort to keep things feeling real for each new person,” added Tanaka.

These findings extend far beyond the lab. In virtual workspaces, therapy sessions, and customer service interactions, avatars may unintentionally nudge users toward riskier behavior. This could be a double-edged sword – boosting confidence in some cases while increasing impulsivity in others.

Next steps in avatar research

Future investigations will likely explore how avatar characteristics – age, gender, realism – affect user responses.

Researchers are also interested in whether these findings apply to other kinds of decision-making, such as moral choices or long-term planning.

The study suggests that more human-like avatars could deepen emotional engagement, potentially altering the balance between empathy and risk even further. As avatar technology becomes more lifelike and widespread, these questions will become increasingly important.

New era of virtual world risk

Tanaka and Haruno’s study offers a timely look at how digital interactions shape human behavior. By revealing that avatars influence our emotional evaluation of uncertainty, they uncover a hidden driver of risk in the virtual world.

The amygdala, a structure tied to emotional memory and fear, emerges as a key player in how we navigate the unfamiliar terrain of avatar communication.

As we spend more time in these environments, we must learn how these subtle shifts in perception can shape the choices we make – online and off.

The study is published in the journal PLOS Biology.

—–

Like what you read? Subscribe to our newsletter for engaging articles, exclusive content, and the latest updates. 

Check us out on EarthSnap, a free app brought to you by Eric Ralls and Earth.com.

—–