Connect with us

Hi, what are you looking for?

Artificial Intelligence

OpenAI data reveal millions of users turn to ChatGPT for help with dark topics every week

OpenAI data reveal millions of users turn to ChatGPT for help with dark topics every week

OpenAI has released a report revealing how people are relying on ChatGPT for mental health issues, and show signs of emotional attachment.

In recent news, several reports surrounding ChatGPT’s role in managing mental health queries have come forward, raising concerns about how individuals are relying on AI chatbots. To address the growing concerns, OpenAI has released a new report revealing how millions of users are using ChatGPT. Shockingly, the majority of users are relying on ChatGPT to have conversations about mental health issues every week, and are having suicidal thoughts.

In addition, several ChatGPT users have an emotional attachment to the chatbot. Furthermore, the company highlights what safety improvements and measures it’s taking to provide the right help to ChatGPT users. OpenAI said, “around 0.15% of users active in a given week have conversations that include explicit indicators of potential suicidal planning.” This counts to more than a million users a week, considering ChatGPT has over 800 million weekly active users.

Thousands of users also show signs of psychosis or mania. With the data, OpenAI further revealed that it has worked with over 170 mental health experts to refine its current model to respond to queries concerning mental health issues. When it comes to the user’s emotional attachment and sensitive topics, the new GPT‑5 model is said to offer 42% fewer problematic responses in comparison to GPT-4o. The report highlighted, “ChatGPT responds more appropriately and consistently than earlier versions.”

These concerns came to attention when a 16-year-old boy shared having suicidal thoughts with ChatGPT, eventually leading the boy to take a big step. Now, OpenAI is not only ordered to protect young people, but also individuals who are going through mental health issues. While the GPT-5 model has claimed to be more effective, the company highlights that there is more to be done as the technology advances. OpenAI said, “We’ll keep advancing both our taxonomies and the technical systems we use to measure and strengthen model behaviour in these and future areas.”

The article originally appeared on Hindustan Times

You May Also Like

World

Pakistan will discuss an Extended Fund Facility (EFF) with the International Monetary Fund (IMF) in Washington next month, Finance Minister Muhammad Aurangzeb said on...

Business

New York CNN — Apple has received approval to change the way its smartwatches function so the company can overcome the Apple Watch ban imposed by...

Science

The Indian Space Research Organisation (ISRO) on Thursday successfully completed the docking process of the SpaDeX satellites. “India docked its name in space history!...

Science

The probability of a “city-killer” asteroid hitting Earth in 2032 just increased to 1 in 32, or 3.1%, according to NASA, and a chilling...