In recent years, generative artificial intelligence has become increasingly present in our daily lives, transforming the way we search for information, compare products, and make decisions.
Tools such as ChatGPT by OpenAI, capable of generating text and suggestions in just a few seconds, represent a significant acceleration of an already ongoing technological process.
But what really happens in consumers’ minds when they interact with AI? Do they trust it? Do they become emotionally involved? Do they pay attention to content in the same way they would with a human interlocutor?
These questions are explored in a study published in the volume Advances in Digital Marketing and eCommerce by Springer (Capone, V., Bartoli, C., Mattiacci, A., Cherubino, P., 2024), which applied Neuromarketing tools to analyze consumer reactions during interactions with ChatGPT.
How the study was conducted
The research combined several neuroscientific and traditional methodologies:
- Eye tracking, to measure visual attention while reading AI-generated responses.
- Facial coding, to detect emotional reactions through facial micro-expressions.
- A final questionnaire, to collect participants’ conscious evaluations regarding usefulness, trust, ease of use, and preferences.
Participants were asked to observe two different interactions with ChatGPT related to a request for advice on purchasing trekking shoes. The AI responses were designed in two different styles:
- Informative and objective answers, focused on technical characteristics and selection criteria.
- More personal and recommendation-oriented answers, with a more consultative tone.
What really captured attention?
Eye-tracking data revealed an interesting result: the more Informative and objective answers attracted more visual attention than the more personal ones. On average, participants spent more time reading descriptive and technical sections, while more personalized suggestions were observed more quickly.
This suggests that, when interacting with AI, consumers tend to look primarily for clear and concrete information, rather than relational engagement.
And from an emotional point of view?
Facial coding analyses showed low levels of emotions such as disgust, fear, joy, surprise, or perplexity, along with a predominance of emotional neutrality.
In other words, interaction with the information provided by ChatGPT generated moderate emotional engagement, without significant peaks either positive or negative. The AI therefore appears to be perceived mainly as a functional tool: useful, but not particularly emotionally stimulating.
What do consumers consciously think?
The questionnaire results confirm this interpretation. Participants reported that:
- AI is perceived as clear and easy to use.
- It is considered useful in everyday life, especially for quickly obtaining information about products and services.
However, a significant portion of participants stated that they still prefer human interaction, particularly when it comes to receiving personalized advice or making more complex decisions.
So (for now), AI yes — but not in place of humans
