One of the most debated topics in artificial intelligence is its apparent ability to show empathy. But… does it truly understand our emotions? Or does it simply simulate them?
How to recognize emotions
Models like GPT-4o have been trained on millions of examples of human interaction. This allows them to identify language patterns associated with emotional states: sadness, enthusiasm, anxiety, gratitude, and so on.
When you write:
“I’m exhausted and I feel like I can’t keep up with everything.”
GPT will likely respond with a supportive tone such as:
"I'm sorry you feel this way. Would you like me to help you organize your tasks to lighten your load?"
There is no real emotional “understanding”, but there is a contextual response based on linguistic probabilities.
Why is this useful?
-
Tactful customer service : many companies already use IAG models that adjust the emotional tone according to the user's state.
-
Initial psychological support : although it does not replace professionals, it can offer a first layer of help.
-
Better interaction in products : assistants like Copilot, Claude or ChatGPT already personalize their response depending on the tone with which you interact.
Risks
-
Excessive anthropomorphism : users may think that the AI actually “feels”.
-
Emotional manipulation : some companies might use this to influence decisions without transparency.
👉 If you're interested in ChatGPT, we recommend reading our article abouttokens in ChatGPT. .