Bias and Misinterpretation

[Jahangir307]

Emotion AI is far from perfect. Cultural, gender, and racial biases are embedded in many training datasets. A smile may mean different things across cultures. Voice pitch may not always indicate anger. Mislabeling a user’s emotional state can lead to incorrect conclusions—and potentially harmful actions.

Moreover, people with neurodivergent traits (e.g., autism) or mental health conditions may not express emotions in stereotypical ways, leading to biased assessments.

4. Case Study: Coca-Cola’s Mood-Based Billboard Campaign
In 2023, Coca-Cola partnered with emotion analytics company vietnam phone number list Realeyes for an interactive ad campaign in London. Digital billboards equipped with facial recognition software scanned pedestrians’ faces in real time, analyzing their expressions to gauge their emotional state.

Depending on the detected emotion—happiness, surprise, boredom, sadness—the billboard displayed a different Coca-Cola product or message. A bored expression triggered a “refresh your moment” ad with Coke Zero. A smile brought up a joyful, high-energy Coca-Cola Classic message.