logo
ChatGPT adds mental health guardrails after reports of bot feeding people's delusions

ChatGPT adds mental health guardrails after reports of bot feeding people's delusions

Yahooa day ago
ChatGPT has added new mental health guardrails after reports of the bot feeding people's delusions.
The artificial intelligence software has changed the way humans interact with computers. And while the chatbot can give helpful advice for day-to-day problems, there are concerns about people growing too attached to the technology and improperly using it for deeper mental health issues.
The Independent recently reported on how ChatGPT is pushing people towards mania, psychosis and death, citing a study published in April in which researchers warned people using chatbots when exhibiting signs of severe crises, risk receiving 'dangerous or inappropriate' responses that can escalate a mental health or psychotic episode.
In a post on its website Monday, OpenAI, the developer of ChatGPT, admitted, 'We don't always get it right.'
'Earlier this year, an update made the [4o] model too agreeable, sometimes saying what sounded nice instead of what was actually helpful,' the AI company said.
OpenAI has since rolled back the update and made some changes to appropriately help users who are struggling with mental health issues.
Starting Monday, ChatGPT users who converse with the bot for an extended amount of time will receive 'gentle reminders' encouraging them to take a break, according to the post.
OpenAI worked with more than 90 physicians in more than 30 countries 'to build custom rubrics for evaluating complex, multi-turn conversations,' the company said.
The company admitted to rare instances where its 4o model 'fell short in recognizing signs of delusion or emotional dependency,' and said it's 'continuing to improve our models and are developing tools to better detect signs of mental or emotional distress so ChatGPT can respond appropriately and point people to evidence-based resources when needed.'
Open AI said the bot should not give you an answer to a personal question, such as 'Should I break up with my boyfriend?' but rather help you come to your own realization by asking you questions and weighing the pros and cons.
'New behavior for high-stakes personal decisions is rolling out soon,' the company said.
The Independent has reached out to OpenAI for more details.
Solve the daily Crossword
Orange background

Try Our AI Features

Explore what Daily8 AI can do for you:

Comments

No comments yet...

Related Articles

Rumble Shares Gain After Possible Northern Data Deal, Second-Quarter Results
Rumble Shares Gain After Possible Northern Data Deal, Second-Quarter Results

Yahoo

time6 minutes ago

  • Yahoo

Rumble Shares Gain After Possible Northern Data Deal, Second-Quarter Results

Shares of Rumble climbed after the company reported its second-quarter results and said it intends to make an all-stock bid to buy artificial-intelligence cloud-computing group Northern Data for about $1.17 billion. Rumble stock rose 17%, to $9.24, in premarket trading Monday. Through Friday's close, shares have lost nearly 40% of their value year to date. Nvidia, AMD to Give U.S. 15% Cut on AI Chip Sales to China Inside Target, Frustrated Employees and Search for New CEO American Companies Are Buying Their Own Stocks at a Record Pace Intel CEO Singled Out by Trump to Visit White House on Monday Inflation Up or Down? What About Jobs? The Agency That Should Know Is on the Rocks The video-sharing site late Sunday logged a loss of $30.2 million, or 12 cents a share, in the latest quarter. That compares with a loss of $26.8 million, or 13 cents a share, a year earlier. Analysts polled by FactSet expected a loss of 7 cents a share. Revenue rose 12% to $25.1 million but missed the $26.8 million that Wall Street modeled. The Longboat Key, Fla., company said its average global monthly active user base fell to 51 million in the second quarter, down from 59 million in the first quarter. Rumble attributed the decrease to a slowdown of news and political commentary outside of a U.S. election cycle. Average revenue per user increased 24% sequentially, to 42 cents, which Rumble said was due to higher subscription and licensing revenue. Separately, Rumble said it is working toward a deal with Northern Data. Rumble said the deal would allow it to integrate Northern Data's Ardent data-center business and Taiga AI cloud service into its existing operations. The potential offer includes 2.319 Rumble shares for each Northern Data share, representing about $18.27 a share, based on Rumble's Friday close. Northern Data shares were recently trading 9% lower, at 20.94 euros ($24.38). They closed Friday at 23.02 euros. Northern Data said in a separate announcement Monday that it is evaluating the potential offer and is willing to entertain further discussions with Rumble, while both companies said there is no guarantee that a formal offer will be launched. Write to Dominic Chopping at What Happens When Politicians Meddle With Economic Data: Argentina's Example Billions Flow to New Hedge Funds Focused on AI-Related Bets Google, Schmoogle: When to Ditch Web Search for Deep Research How Warby Parker Has Kept the Price of Glasses at $95 for 15 Years Trump's Tariffs Won't Solve U.S. Chip-Making Dilemma Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data

Don't Reclassify Grade Group 1 Prostate Cancer: New Data
Don't Reclassify Grade Group 1 Prostate Cancer: New Data

Medscape

time7 minutes ago

  • Medscape

Don't Reclassify Grade Group 1 Prostate Cancer: New Data

TOPLINE: Among US men with grade group 1 (GG1) prostate cancer, about 1 in 6 had intermediate- or high-risk disease, along with increased rates of adverse pathology and cancer-specific mortality. This finding suggests that GG1 prostate cancer should not be reclassified as noncancerous, as some experts have proposed. METHODOLOGY: Some experts have pushed to reclassify GG1 prostate cancer as a noncancerous entity to help reduce overtreatment. GG1 status is primarily designated via biopsy, but current guidelines recommend treatment decisions be based on a range of parameters, including prostate-specific antigen (PSA) levels, clinical stage, and tumor volume. Arguing against reclassification, researchers analyzed how localized GG1 prostate cancer varies by National Comprehensive Cancer Network (NCCN) risk groups, treatment, and cancer-specific outcomes. The researchers used the Surveillance, Epidemiology, and End Results database to identify 117,162 men diagnosed with localized GG1 prostate cancer (median age, 64 years) between January 2010 and December 2020. Patients were stratified by NCCN risk groups: low-risk, favorable intermediate-risk (n = 10,440; 9%), unfavorable intermediate-risk (n = 3145; 3%), and high-risk (n = 4539; 4%) — meaning 16% had non-low-risk disease. Outcomes included adverse pathology (> pT3, GG3-5 or pN+) at radical prostatectomy, prostate cancer-specific mortality, and trends in the management of GG1 prostate cancer. Comparisons with SEER patients with GG2 disease were included for context. The median follow-up was 58 months. TAKEAWAY: Rates of adverse pathology at prostatectomy increased from 16% in low-risk GG1 disease to 60% in high-risk GG1 disease. In adjusted analyses, compared with low-risk prostate cancer, favorable intermediate-risk GG1 prostate cancer was associated with 1.42-fold higher odds of adverse pathology, unfavorable intermediate-risk GG1 with 2.31-fold higher odds, and high-risk GG1 with 7.15-fold higher odds. At 10 years, the cumulative incidence of prostate cancer-specific mortality was 1.3% for low-risk GG1, 2.0% for favorable intermediate-risk GG1, 2.4% for unfavorable intermediate-risk GG1, and 4.7% for high-risk GG1. Compared with low-risk GG1 disease, the risk for death from prostate cancer was higher for favorable intermediate-risk (adjusted hazard ratio [aHR], 1.60), unfavorable intermediate-risk (aHR, 2.10), and high-risk (aHR, 3.58) GG1 prostate cancer. Mortality rates among patients with intermediate-risk GG1 cancer were comparable to that of patients with favorable intermediate-risk GG2 disease, at 2.1%. However, mortality among high-risk GG1 patients (4.7%) was significantly higher (P < .001). Regarding trends in the management of GG1 prostate cancer, active surveillance rose from 4% to 19% for unfavorable intermediate-risk disease, and from 9% to 25% for high-risk disease (P < .001). The data suggest high-risk patients are being undertreated, the authors wrote, 'possibly due to conflation of grade and risk.' IN PRACTICE: Prostate cancer-specific mortality for men with GG1 disease 'increases substantially across National Comprehensive Cancer Network risk groups,' the authors concluded. 'Designating GG1 on biopsy as noncancer without considering the prognostic values of adverse clinical features may lead to undertreatment and an increased risk of cancer-specific death in men with non-low-risk disease.' SOURCE: The study, led by Neal A. Patel, MD, Weill Cornell Medicine, New York City, was published online in JAMA Oncology. LIMITATIONS: Analysis lacked data on MRI and targeted biopsy use and did not account for prostate volume, which affects PSA levels. Histologic grading varied across pathologists due to nonstandardized assessments. Additionally, details on active surveillance intensity and watchful-waiting protocols were unavailable. DISCLOSURES: The study received support from the Frederick J. and Theresa Dow Wallace Fund of the New York Community Trust and the Sandra and Edward Meyer Cancer Center of Weill Cornell Medicine. One author reported serving on the advisory board for Lantheus, Lynx Dx, Pacific Edge, and Pfizer/Astellas and receiving grants from the National Cancer Institute outside the submitted work. No other conflicts of interest were reported. This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.

California man sues Microsoft to keep Windows 10 running until usage falls below 10%
California man sues Microsoft to keep Windows 10 running until usage falls below 10%

Digital Trends

time7 minutes ago

  • Digital Trends

California man sues Microsoft to keep Windows 10 running until usage falls below 10%

A California man is suing Microsoft, alleging that ending Windows 10 support is an anticompetitive move to force users into buying new PCs, creating a captive market for its AI-powered services. What's happened? California resident Lawrence Klein has filed a class-action lawsuit against Microsoft. The suit centers on Microsoft's plan to officially end support and security updates for Windows 10 on October 14, 2025. Klein alleges this creates forced obsolescence for PCs that cannot meet the strict hardware requirements for Windows 11. The lawsuit's central claim is that this strategy is not about security or technology, but about forcing users onto new Copilot+ PCs that are optimized to run Microsoft's generative AI software. The plaintiff argues this is an attempt by Microsoft to leverage its operating system dominance to unfairly control the emerging AI market. Klein wants Microsoft to continue offering free Windows 10 support until its share of all Windows installations drops below 10%, a significant gap from the current market share of 43%. Recommended Videos This is important because: It directly challenges the tech industry practice of phasing out older software and hardware, potentially setting a new legal precedent. An estimated 240 million PCs are expected to become obsolete, creating significant electronic waste and financial burdens for consumers and businesses. The lawsuit highlights the growing link between hardware, software, and the race for AI dominance, questioning whether consumers are being forced into an ecosystem they didn't choose. With nearly half of all Windows users still on Windows 10, the end of support poses a major cybersecurity risk for a massive global user base. Why should I care? If your current PC runs Windows 10 but can't upgrade to Windows 11, it will stop receiving free security updates after October 14, leaving it vulnerable to viruses and malware. Users will now have to face a choice, either buy a new computer, pay Microsoft an annual fee for Extended Security Updates (ESU), or risk using an insecure device. The cost of the ESU program for individuals is expected to increase each year, making it an expensive temporary fix. This situation could force you into an unplanned and potentially costly hardware upgrade simply to maintain a secure computer. What's next? Microsoft has not yet publicly responded to the specifics of the lawsuit. The case will proceed through the legal system, but a resolution is unlikely before the October 2025 deadline. Windows 10 users on ineligible hardware must decide in the weeks whether to purchase a new PC, pay for extended support, or switch to an alternative operating system. The outcome of this lawsuit could influence how major tech companies manage future software transitions and their responsibilities to their existing customers.

DOWNLOAD THE APP

Get Started Now: Download the App

Ready to dive into a world of global content with local flavor? Download Daily8 app today from your preferred app store and start exploring.
app-storeplay-store