Facial Expressions as Emotional Indicators
For a long time, people believed that facial expressions were a reliable indicator of a person’s feelings. However, recent research shows that attempts to recognize emotions based solely on facial expressions almost always fail.
When interacting with others, we instinctively observe their facial expressions and features, drawing conclusions—often subconsciously—about their mood and character. This is how we identify basic emotions like sadness, anger, and joy, and even subtle shades of laughter.
There is a common belief that facial features can reveal certain personality traits, such as selfishness or loyalty, and even indicate illnesses. But how justified is this “face-ism”? The answer remains unclear.
Research: Facial Expressions Alone Are Not Enough
A recent study led by Aleix Martinez from The Ohio State University found that even emotions considered straightforward cannot be clearly read from facial expressions alone. The research team measured the kinetics of facial muscle movements in participants and used algorithms to determine if these patterns could reveal their emotions.
“We wanted to find out if it’s really possible to determine emotions just from facial expressions,” Martinez explains. “And the answer is: no, you can’t.” In their experiments, attempts to read emotions from facial expressions almost always failed. “Everyone changes their facial expressions depending on the situation and cultural background. Not everyone who laughs is happy. And not every happy person laughs.”
AI Algorithms for Emotion Analysis Fall Short
The problem is that some companies have already started using special AI algorithms to interpret people’s facial expressions. These artificial intelligence systems are used, for example, to analyze job candidates. Through video recordings, the system assigns certain emotions and personality traits to faces. Sometimes, cameras even monitor employees at work to determine their emotional state.
Martinez and his team are critical of this trend. After testing several such technologies, they found the results to be completely inadequate. The researchers emphasize that it’s important to understand these systems will never reach 100% accuracy. Algorithms tend to make mistakes almost as often as humans do, especially since they don’t yet consider other factors like skin color, posture, and context.
It’s Not Just About Facial Expressions
Based on their findings, Martinez and his team concluded that recognizing emotions correctly requires more than just facial expressions. Another indicator can be skin color. Emotions cause the brain to release peptides, which change blood flow and composition, and as a result, the color of the face changes.
People also provide other clues, such as their posture. But the most crucial factor is context. In one experiment, Martinez showed participants a fragment of an image showing only a man’s face. His mouth was open as if shouting, and his face was bright red. “When people saw just the fragment, they thought, ‘Wow, this guy is really angry,’” Martinez says. “But when participants saw the full picture, they realized it was a soccer player celebrating a goal. In that context, it was clear he was just happy. But if you look at his face in isolation, he looked almost dangerous.”
Another risk, according to Martinez, is that we might miss someone’s true emotions and misjudge their abilities. “Some people confidently claim they can tell if someone is guilty of a crime, if a student is paying attention in class, or if a customer is satisfied with a purchase just by looking at their face. Our research shows these claims are not true. You simply can’t determine such things this way.”
For example, Martinez mentions a teacher who thinks a student is distracted. The teacher assumes the student is listening only if they smile and nod. But, as Martinez argues, a person can listen attentively without smiling for various reasons. Punishing a student based on this assumption is wrong. The researchers note that everyone—from ordinary people to professors and legal experts—should consider more than just facial expressions before making judgments about others.