Training AI on stereotypical facial expressions is bound to mislead because scientific evidence indicates that there are no universal expressions of emotion
Training algorithms on stereotypical facial expressions is bound to mislead. — Imagine that you are interviewing for a job. X: @carnage4life , @naimulkhan , and @lfeldmanbarrett LinkedIn: Dr. Jeffrey Funk , Rafael Brown , and Charles L Mauro X: Dare Obasanjo / @carnage4life : Every AI system that claims to either read people's emotions or other characteristics like their political affiliations from facial expressions is a scam. There's no underlying science here so there's no underlying “good data” to have trained an AI on to predict these qualities [image] Naimul Khan / @naimulkhan : Forget about face, my lab has been at it for 6+ years and we can't get past 72% even with physiological signals (ECG/EEG). Emotion recognition research has a long way to go, and should NOT be released in the wild like this irresponsibly. https://www.wsj.com/... Lisa Feldman Barrett / @lfeldmanbarrett : If you're interested to view the generative AI output I mention in my @WSJ op-ed, visit http://www.how-emotions-are-made.com/ .... LinkedIn: Dr. Jeffrey Funk : Despite OpenAI's claims, “The best available scientific evidence indicates that there are no universal expressions of emotion.” … Rafael Brown : Microsoft walked away from emotion detection and facial recognition as did IBM and Amazon. Here is a reading list: … Charles L Mauro : Dr. Jeffrey Funk My research team utilizes Micro Facial Expression Analysis research tools that do this exact set of functions. …
Context & Ripple Effects
This is part of a long-running challenge to emotion-recognition systems: earlier coverage found researchers questioning both their efficacy and ethics, with facial expressions only loosely matching reported emotion in studies researchers questioning emotion-recognition accuracy and ethics.
The issue matters because the critique targets the training premise, not merely model performance. It follows earlier warnings that a commercially significant emotion-detection market rested on contested science warnings about the field's disputed scientific foundation.
First-order effects
- Developers and buyers of facial emotion-analysis tools face a direct validity problem: training on stereotyped expressions may produce confident outputs without a reliable scientific mapping from face to emotion.
- Microsoft, IBM, and Amazon's reported moves away from emotion-detection or certain facial-recognition efforts look more consequential as the core inference claim is challenged.
Second-order effects
- Vendors using facial signals in hiring, surveillance, or customer analysis will face greater pressure to substantiate what their systems infer, rather than point only to benchmark accuracy.
- Teams may shift toward narrower, explicitly observable facial-analysis functions or add other signals, but the cited difficulty reaching strong accuracy even with ECG/EEG underscores that additional data does not automatically validate emotion labels.
Third-order effects
- If the field cannot establish a generalizable relationship between expressions and emotions, emotion AI may increasingly be evaluated as a high-risk inference product rather than a standard computer-vision capability.
- The durable divide will be between systems that detect visible features and systems that claim to infer internal states or personal traits; the latter face a substantially higher evidentiary bar.
The trend: AI scrutiny is moving beyond whether models can classify inputs toward whether the human attributes their labels claim to measure have a sound scientific basis.