Why AI's "Empathy Problem" Isn't Actually a Problem
Natan Voitenkov
Jan 17, 2025
•
3
min read
The TL;DR: Ipsos researchers are worried that AI can't do qualitative research well because it lacks empathy. We - Genway AI - think they're missing something huge. Let us explain...
First, some context: late last year, Ipsos published this study about AI moderators (aka AI Agents) in market research. Their concern? AI does not have the "emotional intelligence" of human moderators.
But what if that's actually a feature, not a bug? 🤔 Let's break this down
The Interviewer Effect is Real
The Interviewer Effect isn't just a minor inconvenience – it's a well-documented phenomenon in human-moderated research. We are dealing with a psychological dynamic where participants, faced with a friendly, empathetic interviewer, may modify their responses to gain approval. It happens at both a conscious and unconscious level: participants might choose "better" answers, soften criticisms, and align with what they perceive as the moderator's values. While we are just beginning to study how AI moderation might affect this dynamic, it's worth considering whether reduced social pressure in AI-led interviews could lead to more candid responses.
Emotional Consistency > Emotional Intelligence
Think about it:
Human moderators get tired
They have good days and bad days
They bring their own biases
They're literally different people
Meanwhile, AI just... works. The same way. Every time. For hundreds of interviews. Unless you are using a low-fidelity / off the shelf solution in which case hallucination by the AI Agents is going to have the opposite effect.
The Numbers Don't Lie
Here's the wild part: If AI can deliver 80% of the insight quality at 20% of the cost, we're not talking about a compromise – we're talking about a revolution in accessibility. Suddenly, everyone from startups and small businesses to Forbes 100 enterprises can afford rigorous qualitative research at scale.
But here's what really gets us excited: we are arguing about AI's current capabilities when we should be looking at the vector of improvement.
The AI vs human empathy gap isn't static – it's shrinking. Fast. By the time you finish reading this, AI has probably gotten better at emotional intelligence. Genway AI is evolving at an incredible pace, and we’ll share a lot more in the coming weeks.
The Future is Hybrid
Look, we are not saying we should fire all human moderators (sorry, clickbait headlines!). We simply need to be more thoughtful about when we use each tool and adapt our approach as AI evolves. At this point 👇🏻
Need rich, synthesized data across hundreds of interviews? AI.
Exploring deeply personal topics? Human. (though that’s debatable and a topic for another post.)
Quick user feedback? AI.
Complex emotional territory? Human.
You get the idea.
---
What do you think? Have you tried AI moderation in your research? Let us know – and if you would like to take Genway AI for a spin, visit genway.ai and open a free trial account.
More insights
Generate insights with depth and scale using AI Interviewers