Can AI Read Your Emotions?

This isn't sci fi. This is happening in real life.

285.

You aren't just watching their ads. In testing panels, paid viewers are being watched back, with AI reading their faces through webcams to score every reaction. Vendors like Realeyes sell this facial coding technology to major brands by claiming it predicts exactly how an ad makes you feel. They argue that this data is more accurate than traditional surveys, and marketers absolutely love the promise of knowing which creative will convert before they spend a dime.

Does the science back the claim?

However, there is a massive scientific flaw hidden behind these confident sales pitches. A comprehensive 2019 review of over 1,000 studies by psychologist Lisa Feldman Barrett found that faces alone do not reliably reveal our true feelings (Barrett et al., 2019). People often smile when they are furious or scowl when they are deeply focused. To be fair, Barrett doesn't say that faces do not have signals at all; she's saying that face-only signals are not reliable which is what these vendors are selling. Another fascinating experiment showed that viewers could not even tell winning athletes from losing ones just by looking at their isolated faces (Aviezer, Trope & Todorov, 2012). Context is everything, and relying solely on a webcam to decode human emotion is a fundamentally flawed strategy.

Beyond the shaky science, marketers need to pay close attention to the incoming wave of privacy regulations that are being tackled in workplaces and schools. The European Union has already moved to ban emotion recognition in workplaces, and California recently passed laws to restrict employers from using AI surveillance to infer emotional states. As a strategist, your biggest concern should not just be whether this technology actually works; you need to ask where regulation will put this technology next year. You must thoroughly vet any artificial intelligence tool before integrating it into your tech stack.

Before you buy into the hype of emotion reading software, take a step back and look at the bigger picture. We have to remember that human behavior is far too complex to be reduced to an algorithm or a simple facial scan. Have you seen any brands using AI in ways that feel a little too invasive? Drop your thoughts in the comments below so we can discuss where the line should be drawn!

References

1. Barrett, L. F., Adolphs, R., Marsella, S., Martinez, A. M., & Pollak, S. D. (2019). Emotional expressions reconsidered: Challenges to inferring emotion from human facial movements. Psychological Science in the Public Interest, 20(1), 1–68. https://doi.org/10.1177/1529100619832930

2. Aviezer, H., Trope, Y., & Todorov, A. (2012). Body cues, not facial expressions, discriminate between intense positive and negative emotions. Science, 338(6111), 1225–1230. https://doi.org/10.1126/science.1224313

3. Research Live. (2017, March 22). Emotion measurement predicts sales lift with 75% accuracy. Market Research Society. https://www.Research-Live.com/article/news/emotion-measurement-predicts-sales-lift-with-75-accuracy/id/5020232

4. Campaign. (n.d.). Emotion-reading AI company used by Coca-Cola and Mars raises £12m in funding. https://www.campaignlive.com/article/emotion-reading-ai-company-used-cola-cola-mars-raises-12m-funding/1465304

5. Ogletree Deakins. (2026, September 30). California governor signs 3 bills targeting AI and workplace surveillance. https://ogletree.com/insights-resources/blog-posts/california-governor-signs-3-bills-targeting-ai-and-workplace-surveillance/

6. Bloomberg Law. (2026, September 30). California bans using AI to monitor workers' brains, emotions. https://news.bloomberglaw.com/daily-labor-report/newsom-signs-ban-on-using-ai-to-monitor-worker-brain-activity

- Kristina Bozanich - I am a fractional CMO, media producer, Congressional Record Honoree, and LGBTQ+ Advocate. To see more updates, follow on social media or sign up for email subscriptions.