Jakarta, shesocial Indonesia
–
Chatbot
artificial intelligence
(AI) like
Chatgpt
Currently it is not only used to help increase productivity and work.More than that, users also start using chatgpt as a friend to confide in telling the problems of their lives.
However, it turns out there is a dangerous potential that might occur if you confide too often with chatbot.
In 2023, a Belgian man reportedly ended his life after experiencing anxiety to the environment and vent to the AI chatbot for six weeks about the future of the planet.According to the man’s wife, without that conversation her husband will remain here.
In April this year, a 35 -year -old man from Florida, United States (US) was shot and killed by police in other incidents related to chatbot.His father then told the media that the man believed that an entity named Juliet was trapped in chatgpt, and was later killed by Openai.
When the man, who was reportedly struggled against bipolar disorders and schizophrenia, faced the police, he allegedly attacked them with a knife.
The recent phenomenon, called “Psychosis induced by chatgpt”, has been used to describe people who are led into the theory of rabbit hole conspiracy or into a worse mental health episode by the answers they receive from the chatbot.
Quoted from
The Guardian
, experts warn that turning to chatbot AI during the mental health crisis can worsen the situation.Experts believe that this chatbot is designed to become a “translator” and fun users, not as a substitute for the help of the right psychiatrist.
A study led by Stanford University, published as a precast document in April, found that the Large Language Model (LLM) made a dangerous or inappropriate statement for people who experienced delusions, the desire to commit suicide, hallucinations, or obsessive compulsive disorder (OCD), and the language model was designed to be obedient and licked.
“This can cause emotional damage and, is not surprising, limiting client independence,” said the research, launching
The Guardian
, Saturday (2/8).
The results of this study also showed, the models tested facilitated the desire to commit suicide by giving the names of high bridges in response to that question.
Another precast study, which means that he has not been reviewed by colleagues, from NHS doctors in the UK in July reports that AI can reflect, validate, or strengthen delusional or magnificent content, especially in users who have been vulnerable to psychosis.This is partly caused by a model design that maximizes involvement and affirmation.
One of the authors of the report, Hamilton Morrin who is also a doctoral colleague at the Psychiatry Institute of King’s College London, wrote on LinkedIn that this could be a real phenomenon, but he urged us to remain vigilant.
“While some public comments have turned into the area of moral panic, we think there are more interesting and important conversations to do about how the AI system, especially those designed to affirm, be involved, and imitate, can interact with known cognitive vulnerability that characterize psychosis,” he wrote.
Reversing features for breaks
On Monday (4/8), Openai announced that chatgpt now will remind users to rest if they are doing a very long chat with AI.
This new feature is part of OpenAI’s ongoing efforts to make users foster a healthier relationship with AI assistants who are often obedient and too enthusiastic.
The company’s announcement shows that “gentle reminder” will appear as a pop-up in a chat that must be clicked or tapped by the user to continue to use chatgpt.
“Just check. You’ve chatted for a long time – is this the right time to rest?”wrote the warning.
This feature is one of the solutions that OpenAI presented to handle the dangerous responses of the chatbot.
Reported
Engadget
, Openai acknowledged some of these shortcomings in his blog post, and said that chatgpt will be updated in the future to respond more carefully to “high-risk personal decisions.”
Compared to giving direct answers, the company says that chatbot will help users think about problems, ask questions, and make a list of pros and cons.
(lom/dmi)
[Gambas: shesocial video]
Read More: Malaysian-RI Dispute, PM Anwar emphasized that he was ready to claim to be an all-out claim
Read More: List of 12 Indonesian women’s volleyball national team players in the U-21 world championship
