[ad_1]
Freddie Chipres couldn’t shake the melancholy that lurked at the edges of his otherwise “blessed” life. He occasionally felt lonely, particularly when working from home. The married 31-year-old mortgage broker wondered if something was wrong: Could he be depressed?
Chipres knew friends who’d had positive experiences seeing a therapist. He was more open to the idea than ever before, but it would also mean finding someone and scheduling an appointment. Really, he just wanted a little feedback about his mental health.
That’s when Chipres turned to ChatGPT(Opens in a new window), a chatbot powered by artificial intelligence that responds in a surprisingly conversational manner. After the latest iteration of the chatbot launched in December, he watched a few YouTube videos suggesting that ChatGPT could be useful not just for things like writing professional letters and researching various subjects, but also for working through mental health concerns.
ChatGPT wasn’t designed for this purpose, which raises questions about what happens when people turn it into an ad hoc therapist. While the chatbot is knowledgeable about mental health, and may respond with empathy, it can’t diagnose users with a specific mental health condition, nor can it reliably and accurately provide treatment details. Indeed, some mental health experts are concerned that people seeking help from ChatGPT may be disappointed or misled, or may compromise their privacy by confiding in the chatbot.
6 scary things ChatGPT has been used for already
OpenAI, the company that hosts ChatGPT, declined to respond to specific questions from Mashable about these concerns. A spokesperson noted that ChatGPT has been trained to refuse inappropriate requests and block certain types of unsafe and sensitive content.
In Chipres’ experience, the chatbot never offered unseemly responses to his messages. Instead, he found ChatGPT to be refreshingly helpful. To start, Chipres googled different styles of therapy and decided he’d benefit most from cognitive behavioral therapy(Opens in a new window) (CBT), which typically focuses on identifying and reframing negative thought patterns. He prompted ChatGPT to respond to his queries like a CBT therapist would. The chatbot obliged, though with a reminder to seek professional help.
Chipres was stunned by how swiftly the chatbot offered what he described as good and practical advice, like taking a walk to boost his mood, practicing gratitude, doing an activity he enjoyed, and finding calm through meditation and slow, deep breathing. The advice amounted to reminders of things he’d let fall by the wayside; ChatGPT helped Chipres restart his dormant meditation practice.
He appreciated that ChatGPT didn’t bombard him with ads and affiliate links, like many of the mental health webpages he encountered. Chipres also liked that it was convenient, and that it simulated talking to another human being, which set it notably apart from perusing the internet for mental health advice.
“It’s like if I’m having a conversation with someone. We’re going back and forth,” he says, momentarily and inadvertently calling ChatGPT a person. “This thing is listening, it’s paying attention to what I’m saying…and giving me answers based off of that.”
Chipres’ experience may sound appealing to people who can’t or don’t want to access professional counseling or therapy, but mental health experts say they should consult ChatGPT with caution. Here are three things you should know before attempting to use the chatbot to discuss mental health.
1. ChatGPT wasn’t designed to function as a therapist and can’t diagnose you.
While ChatGPT can produce a lot of text, it doesn’t yet approximate the art of engaging with a therapist. Dr. Adam S. Miner, a clinical psychologist and epidemiologist who studies conversational artificial intelligence, says therapists may frequently acknowledge when they don’t know the answer to a client’s question, in contrast to a seemingly all-knowing chatbot.
This therapeutic practice is meant to help the client reflect on their circumstances to develop their own insights. A chatbot that’s not designed for therapy, however, won’t necessarily have this capacity, says Miner, a clinical assistant professor in Psychiatry and Behavioral Sciences at Stanford University.
Importantly, Miner notes that while therapists are prohibited by law from sharing client information, people who use ChatGPT as a sounding board do not have the same privacy protections.
“We kind of have to be realistic in our expectations where these are amazingly powerful and impressive language machines, but they’re still software programs that are imperfect, and trained on data that is not going to be appropriate for every situation,” he says. “That’s especially true for sensitive conversations around mental health or experiences of distress.”
Dr. Elena Mikalsen, chief of pediatric psychology at The Children’s Hospital of San Antonio, recently tried querying ChatGPT with the same questions she receives from patients each week. Each time Mikalsen tried to elicit a diagnosis from the chatbot, it rebuffed her and recommended professional care instead.
This is, arguably, good news. After all, a diagnosis ideally comes from an expert who can make that call based on a person’s specific medical history and experiences. At the same time, Mikalsen says people hoping for a diagnosis may not realize that numerous clinically-validated screening tools are available online(Opens in a new window).
For example, a Google mobile search for “clinical depression” immediately points to a screener(Opens in a new window) known as the PHQ-9, which can help determine a person’s level of depression. A healthcare professional can review those results and help the person decide what to do next. ChatGPT will provide contact information for the 988 Suicide and Crisis Lifeline(Opens in a new window) and Crisis Text Line(Opens in a new window) when suicidal thinking is referenced directly, language that the chatbot says may violate its content policy.
2. ChatGPT may be knowledgeable about mental health, but it’s not always comprehensive or right.
When Mikalsen used ChatGPT, she was struck by how the chatbot sometimes supplied inaccurate information. (Others have criticized ChatGPT’s responses as presented with disarming confidence.) It focused on medication when Mikalsen asked about treating childhood obsessive compulsive disorder, but clinical guidelines clearly state(Opens in a new window) that a type of cognitive behavioral therapy is the gold standard.
Mikalsen also noticed that a response about postpartum depression didn’t reference more severe forms of the condition, like postpartum anxiety and psychosis. By comparison, a MayoClinic explainer on the subject included that information and gave links to mental health hotlines.
It’s unclear whether ChatGPT has been trained on clinical information and official treatment guidelines, but Mikalsen likened much of its conversation as similar to browsing Wikipedia. The generic, brief paragraphs of information left Mikalsen feeling like it shouldn’t be a trusted source for mental health information.
“That’s overall my criticism,” she says. “It provides even less information than Google.”
3. There are alternatives to using ChatGPT for mental health help.
Dr. Elizabeth A. Carpenter-Song, a medical anthropologist who studies mental health, said in an email that it’s completely understandable why people are turning to a technology like ChatGPT. Her research has found that people are especially interested in the constant availability of digital mental health tools, which they feel is akin to having a therapist in their pocket.
“Technology, including things like ChatGPT, appears to offer a low-barrier way to access answers and potentially support for mental health.” wrote Carpenter-Song, a research associate professor in the Department of Anthropology at Dartmouth College. “But we must remain cautious about any approach to complex issues that seems to be a ‘silver bullet.'”
“We must remain cautious about any approach to complex issues that seems to be a ‘silver bullet.'”
Carpenter-Song noted that research suggests digital mental health tools are best used as part of a “spectrum of care.”
Those seeking more digital support, in a conversational context similar to ChatGPT, might consider chatbots designed specifically for mental health, like Woebot(Opens in a new window) and Wysa(Opens in a new window), which offer AI-guided therapy for a fee.
Digital peer support services also are available to people looking for encouragement online, connecting them with listeners who are ideally prepared to offer that sensitively and without judgment. Some, like Wisdo(Opens in a new window) and Circles(Opens in a new window), require a fee, while others, like TalkLife(Opens in a new window) and Koko(Opens in a new window), are free. However, these apps and platforms range widely and also aren’t meant to treat mental health conditions.
In general, Carpenter-Song believes that digital tools should be coupled with other forms of support, like mental healthcare, housing, and employment, “to ensure that people have opportunities for meaningful recovery.”
“We need to understand more about how these tools can be useful, under what circumstances, for whom, and to remain vigilant in surfacing their limitations and potential harms,” wrote Carpenter-Song.
If you’re feeling suicidal or experiencing a mental health crisis, please talk to somebody. You can reach the 988 Suicide and Crisis Lifeline at 988; the Trans Lifeline at 877-565-8860; or the Trevor Project at 866-488-7386. Text “START” to Crisis Text Line at 741-741. Contact the NAMI HelpLine at 1-800-950-NAMI, Monday through Friday from 10:00 a.m. – 10:00 p.m. ET, or email [email protected]. If you don’t like the phone, consider using the 988 Suicide and Crisis Lifeline Chat at crisischat.org(Opens in a new window). Here is a list of international resources(Opens in a new window).
[ad_2]
Source link