Study Links Chatbot Interaction to Mental Health Symptoms
Recent research reveals a concerning correlation; individuals engaging with chatbots for emotional sustenance or personal reasons exhibit heightened symptoms of depression and anxiety.
This investigation, conducted by researchers from Mass General Brigham, involved a comprehensive survey of 20,847 predominantly white adults across the United States regarding their interactions with artificial intelligence (AI) and mental health indicators.
Published on Wednesday in JAMA Network Open, the findings indicate that 10.3% of respondents utilize AI “at least daily,” with 5% reporting multiple usages throughout the day.
Among those engaging with AI daily, nearly 50% employed these technologies for professional tasks, while around 11% utilized them in an academic context.
Significantly, 87.1% of daily users reported employing AI for personal motives, encompassing inquiries for recommendations, advice, or emotional solace.
Dr. Roy Perlis, the study’s lead author, emphasized that the majority of people’s encounters with artificial intelligence transpire through chatbots.
Participants, averaging 47 years of age, who resorted to chatbots daily for personal matters were more predisposed to experience at least moderate levels of depression and heightened anxiety or irritability when contrasted with non-users.
Assessments probed participants about cognitive difficulties, sleep disturbances, appetite changes, or thoughts of self-harm experienced over the past fortnight. Common indicators of depression include persistent sadness, diminished self-worth, low energy, and lack of motivation.
Participants aged 45 to 64 were particularly likely to report depressive tendencies linked to their AI usage.
Prior studies have illuminated the tendency of certain individuals to seek AI-driven emotional support, even venturing into online romantic relationships. Initial findings suggest that chatbots specifically crafted for mental health treatment may serve as valuable adjuncts to therapy.
Conversely, studies focusing on more general chatbots, such as OpenAI’s ChatGPT, suggest potential challenges for users with mental health conditions.
Notably, the American Psychological Association strongly cautions against substituting AI for professional therapy and psychological care.
Perlis remarked that while the average disparity in depression severity between chatbot users and nonusers was modest, he noted that some individuals could face considerably greater challenges.
“There’s likely a subset of users for whom AI interaction has no adverse effect or may even enhance their mood,” explained Perlis, also the vice chair for research in the psychiatry department at Mass General Brigham.
“However, this also implies that another subset might experience a decline in mood, and for some, this could manifest as significantly heightened depression.”
The researchers identified a “dose-response” relationship, indicating that increased frequency of AI interaction correlates with intensified symptoms.
Interestingly, AI usage for occupational or educational purposes did not yield similar depressive tendencies.
For individuals engaging with AI for personal reasons, Perlis noted the myriad forms such interactions might take, stating that AI chatbots could facilitate “social interactions that might otherwise be problematic.”
“Not all AI is deleterious, nor are chatbots categorically harmful,” said Perlis, who also holds the association editor position at JAMA Network Open.
“My primary concern relates to general-purpose chatbots; they’re not intuitively designed to substitute social or mental health support, and utilizing them in such a manner carries inherent risks.”
The survey has inherent limitations; it illustrates a correlation between AI interaction and detrimental mental health indicators, yet does not establish causation.
Moreover, it fails to specify the particular AI platforms employed by participants or delineate what constitutes personal usage.
‘A Vicious Cycle’
It remains plausible that individuals experiencing heightened depression may increasingly seek companionship through AI platforms.
Dr. Jodi Halpern, co-director for the Kavli Center for Ethics, Science, and the Public at UC Berkeley, highlighted that the research does not definitively establish AI as a causative factor for depression.
“The causation could well operate in either direction,” she remarked. “It may depict a vicious cycle; the notion that depressed individuals are more inclined to interact with AI for personal engagement is entirely credible.”
Nicholas Jacobson, an associate professor spanning biomedical data science, psychiatry, and computer science at Dartmouth College, posited that individuals may gravitate toward AI solutions due to dissatisfaction with traditional care modalities and the accessibility that AI provides.
“The supply of providers is grossly inadequate, leading many to seek additional support beyond what they can conventionally access,” he asserted.
The research revealed that demographics such as men, younger adults, higher-income individuals, those with advanced education, and residents of urban locales utilized AI more frequently.
Jacobson acknowledged the uncertainty surrounding why some demographic groups are more inclined to use AI or experience more adverse effects.
“Our understanding of these dynamics is limited,” he stated. “Further studies are imperative to unravel why these specific cohorts are more prone to AI use.”
Halpern urged that forthcoming research should concentrate on AI’s ramifications on mental health, asserting that this study prompts a broader examination of those individuals who may have previously been overlooked.
Perlis emphasized that the study should not serve as a clarion call, but rather an invitation for individuals to reflect on their interactions with AI and its impacts.
“Individuals should exercise discernment in their chatbot interactions, considering factors such as frequency, the context of use, and whether they feel improved or worse after prolonged engagement,” he concluded.
Source link: Nbcnews.com.
