Psychological first aid (PFA) is a set of skills used in the acute period after encountering traumatic situations, aimed at limiting distress and maladaptive behavior. The creation of PFA tools in a digital format increases the accessibility of assistance, creates the possibility of scaling, adherence to a clear methodological sequence.

The goal is to determine the impact of factors such as being in danger, the inability to change the situation, and others, on the stress level of users of the “Friend” PFA chatbot, as well as to evaluate the effectiveness of this chatbot as a means of reducing stress during a certain period of interaction.

The research was conducted on the basis of the psychological first aid chatbot “Friend” in the Telegram messenger in the period from October 2022 to September 2023.

Thus, of all users who interacted with the bot, 54 % passed the first measurement of the stress level, and 9.7 % reached the fourth measurement. 94.7 % (28436) indicated that they were safe when interacting with the bot. Most of those who gave a negative answer to this question answered that they needed support (77.1 %, 1179). Of them, 69.4 % (802) answered that they could not influence the situation due to which they are in danger. 3 % (909) of users indicated that they did not have enough food or water. Almost everyone who continued to interact with the bot 99.6 % (20563) answered that they felt lonely. the highest relative number of new users with severe stress was recorded in July 2023 and October 2023. The lowest number of users who indicated that they do not feel stressed at all was recorded in May 2023. There is a statistically significant difference between the measurements (χ2(3) = 2218, p < 0.001), the stress level decreased from an of 3.89 (Med = 4, St.Dev. = 1.06) at the first measurement to 2.35 (Med = 2, St . Dev. = 1.28) points at the last measurement. The change had a gradual, linear character, where each measurement differs from the other (Durbin-Conover’s p < 0.001 in all pairwise comparisons). According to the criterion of pairwise comparison, the largest change is observed between 1 and 4 measurements (H = 64.8), the smallest between 3 and 4 measurements (H = 11.5), which may indicate a gradual plateauing and a decrease in the effect of interaction with the bot. p < 0,001), рівень стресу зменшився з 3,89 (Med = 4, St.Dev. = 1,06) на першому замірі до 2,35 (Med = 2, St.Dev. = 1,28) балів на останньому замірі. Зміна мала градуальний, лінійний характер, де кожний замір відрізняється від іншого (Durbin-Conover’s p < 0,001 у всіх попарних порівняннях). За критерієм попарного порівняння найбільша зміна спостерігається між першим та четвертим заміром (H = 64,8), найменша — між третім та четвертим замірами (H = 11,5), що може свідчити про поступовий вихід на плато та зменшення ефекту від взаємодії з ботом.

A study of user interaction with a chatbot showed important changes in stress levels. Using statistical methods, significant relationships were found between the level of stress and various factors, such as feelings of safety, need for support, and traumatic experiences. Interaction with the bot significantly reduced the stress level of users when comparing the level of stress before and after using the bot, which serves as a rationale for conducting further research on the effectiveness of this method of support.