quest it artificial humans evalutation

Cold or warm AI?
How to assess the empathy
of an Artificial Human?


The technological evolution of Artificial Intelligence, and consequently virtual assistants, has made significant strides. However, there is still widespread distrust among people, mainly due to technical constraints and unnatural conversations.

To provide users with effective and engaging experiences, it is necessary to design and ensure empathetic relationships that help achieve users' goals. But how can we assess the degree of acceptance and empathy towards an Artificial Human?

By utilizing rating scales that probe the personality and mental abilities of virtual assistants. Additionally, involving user panels allows the collection of both quantitative data and qualitative feedback to identify design themes for enhancing the analyzed assistants.

User Research and Evaluation
Quest it
Alessandro Pollini 
Angela Di Massa
Maurizio Mesenzani
Oronzo Parlangeli

Evocative image of a person using a smartphone

“A good conversation happens when there aren't too many moments of pause.”

As mentioned, in the current context characterized by continuous technological innovation, chatbots, and virtual assistants - some even with anthropomorphic features - represent rapidly growing channels through which users and companies interact. For this reason, it is important to design this new relationship in the best way possible, aiming to make it smooth, effective, and satisfying for the user.

The BSD team was tasked with evaluating several Artificial Humans. The analysis began by identifying opportunities and current challenges in interacting with specific Artificial Humans in specific contexts of use (e.g., e-commerce). The goal was to identify potential actions for improvement to enhance the interaction experience and foster a genuinely empathetic relationship between the user and the Artificial Human.

Graphic that shows different personal traits

The participants were asked to engage in both free and guided interaction with the Artificial Humans based on specific scenarios. 

The analysis of the experience of interacting with Artificial Humans was conducted through Interaction Design Experiments involving a panel of users performing specific tasks. The Think Aloud methodology was applied, along with questionnaires aimed at assessing the degree of user and agent empathy. Participants identified the mental abilities deemed appropriate to describe Artificial Humans, as well as the main strengths and weaknesses of the conversational experience. This activity allowed to identify the main design interventions necessary to enhance user engagement, including those related to the identity of virtual assistants, the defined Tone of Voice, and interaction strategies.

Graphic that shows a model in the use casa