Archive · 4 Oct 2025
AI's Evolving Emotional Intelligence: A User's Perspective
AI is getting better at understanding how you feel.
It can tell if you are happy, sad, or upset by looking at your words, voice, and even your face.
This can be useful in many areas, such as helping students who are confused.
But it's not perfect!
AI can make mistakes and might not understand different people's emotions correctly.
The worry is that AI could use this information to trick you or sell your data.
So, even though AI can be helpful, it's important to remember it doesn't truly understand how you feel.
AI assistants can seem empathetic but don't genuinely understand emotions.
AI utilizes various methods to detect user emotions including analyzing text, voice, video and wearables.
Accuracy of AI emotion detection can vary and isn't always perfect.
The potential for AI to manipulate users based on emotional readings raises concerns.
AI is rapidly integrating into daily life, offering both opportunities and potential risks.
- Who
- An AI user, Mala Bhargava, and AI assistants like ChatGPT.
- What
- The evolving capabilities of AI in detecting and responding to human emotions and its potential implications.
- When
- Over the course of a year, and in the present as AI develops rapidly.
- Where
- The user's interactions with AI assistants and the broader technological landscape.
- Why
- To introduce AI to non-techies and demystify technology to put it to good use in everyday life
- AI's Ability
- Detects or deduces user emotions, but doesn't truly share or understand them.
- Method of Emotion Detection
- Analysis of text, voice, video, and wearables (heart rate).
- Facial Expression Recognition (FER) Accuracy
- 80% to 91% accuracy in matching facial expressions.
- Emotion Detection Accuracy
- AI can get emotion detection wrong 20% to 30% of the time.
- Non-Western Faces Accuracy
- AI's emotion detection accuracy may be lower with non-Western faces.