A simple chat on a computer with a chatbot, the little character that appears on the screen, can prompt some people to get vaccinated against Covid-19. This is the conclusion of a study by CNRS and Inserm.
Reading time : 2 min.
The initial idea is that mass communication on Covid-19, TV ads on the radio, does not work. To convince those who hesitate, nothing beats a discussion with a trusted person, knowledgeable on the issue, ideally their attending physician, who can answer all the questions that arise, from the most personal to the most general. But it’s impossible to offer this nationwide, hence the idea of this chatbot, or conversational agent, the little robot that appears on the screen and asks if he can respond to Questions.
Cognitive science researchers from CNRS, Inserm and the École nationale supérieure have developed a chatbot capable of answering 51 questions, those most often asked about vaccines. For example: are vaccines safe? Are there any side effects, what kind? The answers are short and the interlocutor can move on to another question: “You told me there were no side effects but I heard the opposite.”
The team tested the chatbot on 338 people and compared the result with a control group of 305 people who only had access to a vaccination backgrounder. The number of participants with a positive opinion of vaccines increased by 37% in the group that exchanged several minutes with the robot. More importantly, they were more inclined to be vaccinated, the number of opponents to the vaccination fell by 20% in this group. And that’s not all, the chatbot effect has repercussions beyond the control group since half of the people who spoke with him then tried to convince relatives to be vaccinated and three quarters used the arguments of the chatbot. All these results have just appeared in the journal Journal of Experimental Psychology Applied.
This is what these researchers want. Hugo Mercier works at the CNRS, he thinks that this chatbot could perfectly find its place in the TousAnti Covid application, provided that the information of the little robot is rigorously updated. Because if there is the slightest error in the answers, the whole system loses its credibility.