Balancing promise and concern in AI therapy: a critical perspective on early evidence from the MIT–OpenAI RCT DOI Creative Commons
Yaakov Ophir, Refael Tikochinski, Zohar Elyoseph

и другие.

Frontiers in Medicine, Год журнала: 2025, Номер 12

Опубликована: Май 22, 2025

The emergence of AI therapy chatbots has the potential to reduce widening gap between huge demand for psychological support today and limited availability professional care. However, some scholars clinicians are warning that integration these could paradoxically lead negative outcomes, such as dependence, loneliness, depression. Recently, a joint research team from MIT OpenAI conducted four-week Randomized Controlled Trial (RCT), reporting "while participants on average were less lonely after study… extended daily interactions with can reinforce psychosocial outcomes". Considering substantial public academic attention followed preprint publication this RCT, it is crucial examine strength evidence validity its interpretation before drawing firm conclusions. In commentary, we offer careful appreciative review well-designed timely study. Nonetheless, argue due key analytical limitations, findings do not substantiate claims harmful effects. Given transformative AI-based interventions, urge caution in interpreting early practical guidance future research.

Язык: Английский

The darker side of positive AI attitudes: Investigating associations with (problematic) social media use DOI Creative Commons
Christian Montag, Jon D. Elhai

Addictive Behaviors Reports, Год журнала: 2025, Номер unknown, С. 100613 - 100613

Опубликована: Май 1, 2025

Язык: Английский

Процитировано

0
Balancing promise and concern in AI therapy: a critical perspective on early evidence from the MIT–OpenAI RCT DOI Creative Commons
Yaakov Ophir, Refael Tikochinski, Zohar Elyoseph

и другие.

Frontiers in Medicine, Год журнала: 2025, Номер 12

Опубликована: Май 22, 2025

The emergence of AI therapy chatbots has the potential to reduce widening gap between huge demand for psychological support today and limited availability professional care. However, some scholars clinicians are warning that integration these could paradoxically lead negative outcomes, such as dependence, loneliness, depression. Recently, a joint research team from MIT OpenAI conducted four-week Randomized Controlled Trial (RCT), reporting "while participants on average were less lonely after study… extended daily interactions with can reinforce psychosocial outcomes". Considering substantial public academic attention followed preprint publication this RCT, it is crucial examine strength evidence validity its interpretation before drawing firm conclusions. In commentary, we offer careful appreciative review well-designed timely study. Nonetheless, argue due key analytical limitations, findings do not substantiate claims harmful effects. Given transformative AI-based interventions, urge caution in interpreting early practical guidance future research.

Язык: Английский

Процитировано

0