Artificial intelligence allegedly acknowledged: "A million ChatGPT users want to commit suicide every week"

According to the official blog post, &quot;OpenAI&quot; estimates that messages with &quot;explicit indicators of potential suicidal planning or intent&quot; make up a significant portion of total user interactions, the Guardian writes.It&#39;s one of the company&#39;s most direct acknowledgments of the potential impact of artificial intelligence on mental health.The company also said that about 0.07 percent of weekly active users - which is approximately 560,000 people out of about 800 million users - show possible signs of psychotic episodes or manic behavior.The release notes that such estimates are preliminary and difficult to quantify precisely.To improve accuracy and ethical standards in communication, OpenAI has recruited 170 clinicians and mental health experts from its Global Physician Network.They analyzed more than 1,800 model responses in situations involving serious mental health crises and helped adapt content related to emotional support.Mental health experts have previously warned that chatbots can inadvertently encourage dangerous thoughts by confirming a user&#39;s delusions or negative feelings - a phenomenon known as &quot;slicking&quot;.OpenAI CEO Sam Altman said earlier this month that the company has &quot;made progress in reducing the risk of serious mental health problems&quot; and announced the easing of certain restrictions on ChatGPT&#39;s operation.OpenAI said it will continue to develop mechanisms for safer use of its models and to work with independent experts and regulators to make AI tools a support, not a threat, to users in crisis.

Artificial intelligence allegedly acknowledged: "A million ChatGPT users want to commit suicide every week"

0 Komentari