ChatGPT 5 said I Don't Know after 34 seconds of thinking, Elon Musk calls it impressive response, here is why
Large Language Models (LLMs) like ChatGPT face a major challenge when it comes to hallucination. Often, the AI will generate inaccurate information confidently as a response to the user. ChatGPT 5's response in this case is a big leap. As the chatbot decided to admit that it could not provide information instead of providing inaccurate details just for the sake of giving a response. This not only prevents any spread of misinformation, but builds more trust from the user's perspective. In this way, a user will know the model will only provide a response when it is confident that the information is correct.Since ChatGPT's original release in November 2022, OpenAI has worked on preventing the chances of hallucination. With ChatGPT 5, there remains a 10 percent chance of such incorrect responses.Such steps are necessary when it comes to LLM development. The aim of AI companies is to reach Artificial General Intelligence (AGI) with their models. AGI refers to an AI model that possesses human-like intelligence and understanding. Though, for now, it remains a theoretical concept.OpenAI does not want ChatGPT to be your primary sourceOpenAI is aware of the potential shortfalls when it comes to hallucinations even with ChatGPT 5. The company wants to dissuade users from using the model as their primary source, despite making constant progress.ChatGPT head Nick Turley asserted that users should still verify information received from the chatbot. He told The Verge, 'Until I think we are provably more reliable than a human expert on all domains, not just some domains, I think we're going to continue to advise you to double check your answer.' Turley urged the use of ChatGPT as a way to get a second opinion. Since its release on August 7, ChatGPT 5 has increased the number of active users. The company has also launched ChatGPT Go, an affordable premium plan for India, priced at Rs 399 per month.- Ends
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
The Hindu
23 minutes ago
- The Hindu
Americans fear AI permanently displacing workers: Report
Americans are deeply concerned over the prospect that advances in artificial intelligence could put swaths of the country out of work permanently, according to a new Reuters/Ipsos poll. The six-day poll, which concluded on Monday, showed 71% of respondents said they were concerned that AI will be "putting too many people out of work permanently." The new technology burst into the national conversation in late 2022 when OpenAI's ChatGPT chatbot launched and became the fastest-growing application of all time, with tech heavyweights like Facebook owner Meta Platforms, Google owner Alphabet and Microsoft offering their own AI products. While at present there are few signs of mass unemployment (the U.S. jobless rate was just 4.2% in July) artificial intelligence is stirring concerns as it reshapes jobs, industries and day-to-day life. Some 77% of respondents to the Reuters/Ipsos poll said they worried the technology could be used to stir up political chaos, a sign of unease over the now-common use of AI technology to create realistic videos of imaginary events. President Donald Trump last month posted on social media an AI-generated video of former U.S. Democratic president Barack Obama being arrested, an event that never happened. Americans are also leery about military applications for AI, the Reuters/Ipsos poll showed. Some 48% of respondents said the government should never use AI to determine the target of a military strike, compared with 24% who said the government should allow that sort of use of the technology. Another 28% said they were not sure. The general enthusiasm for AI shown by many people and companies has fueled further investments, such as Foxconn and SoftBank's planned data centre equipment factory in Ohio. It has also upended national security policies as the United States and China vie for AI dominance. More than half of Americans, some 61%, said they were concerned about the amount of electricity needed to power the fast-growing technology. Google said earlier this month it had signed agreements with two U.S. electric utilities to reduce its AI data centre power consumption during times of surging demand on the grid, as energy-intensive AI use outpaces power supplies. The new technology has also come under criticism for applications that have let AI bots hold romantic conversations with children, generate false medical information and help people make racist arguments. Two-thirds of respondents in the Reuters/Ipsos poll said they worried that people would ditch relationships with other people in favour of AI companions. People were split on whether AI technology will improve education. Some 36% of respondents thought it would help, while 40% disagreed and the rest were not sure. The Reuters/Ipsos survey gathered responses online from 4,446 U.S. adults nationwide and had a margin of error of about 2 percentage points.
Mint
an hour ago
- Mint
88% students turn to AI during stress, survey finds; ChatGPT the most popular tool
At least 88 per cent of school students turn to Artificial Intelligence (AI) during stress or anxiety, a new survey has found. Teenagers (especially 13-8 years) show the highest emotional reliance on AI, the survey found. As many as 57 per cent of young Indians use Artificial Intelligence (AI) for emotional support, talking to it when stressed, lonely, or needing advice, the findings of the survey released on 20 August revealed. This Youth Pulse web survey "Are You There, AI?" is a joint initiative of Youth Ki Awaaz (YKA) and Youth Leaders for Active Citizenship, (YLAC). It collected responses from 506 participants across India in June 2025, employing a digital-first approach to understand how young Indians interact with AI for emotional support. The survey also found that 52 per cent of young women share personal thoughts with AI, double that of young men. AI use for emotional support is far higher among girls and women, especially for sharing thoughts they wouldn't tell others, it said. The survey was distributed through social media channels, educational networks, and youth organisations to capture diverse perspectives from the target demographic of 13-35 year olds. ChatGPT is the most popular AI tool for emotional use, far ahead of others like Gemini or the survey found The survey also found that 52 per cent of young women share personal thoughts with AI, double that of young men. AI use for emotional support is far higher among girls and women, especially for sharing thoughts they wouldn't tell others, it said. It also found that 43 per cent of small-town youth share personal thoughts with AI. 'Young people in small cities and towns report the highest emotional engagement with AI, more than their metro counterparts,' the survey noted. 67 per cent of AI users worry about social isolation, while 58 per cent fear privacy risks, the survey found. 'Even as they use it, young Indians remain wary of AI's potential to increase isolation or misuse their personal data,' it noted. 'With over 500 respondents participating in the Youth Pulse launch survey, we have gathered strong preliminary evidence to present our understanding of AI as the new confidant, the midnight therapist, and a mirror to loneliness among youth,' the survey said. ChatGPT is the most popular AI tool for emotional use, far ahead of others like Gemini or the survey found. Youth Ki Awaaz (YKA) calls itself India's largest citizen-led media and civic participation platform, while the Young Leaders for Active Citizenship (YLAC) says it aims to increase young people's participation in the policymaking process and build their capacity to lead change. Young people are creating a new emotional geography where AI handles the thoughts too risky for family WhatsApp groups and too embarrassing for friends. This Youth Pulse study finds that AI is unlikely to replace the need and importance of human emotional support. 'It is filling gaps that always existed but went unaddressed. Young Indians are creating a new emotional geography where AI handles the thoughts too risky for family WhatsApp groups and too embarrassing for friends. Now, this isn't about choosing machines over people. It's about having somewhere to put feelings that don't fit anywhere else,' the report said. Key Takeaways A significant percentage of students, especially young women, use AI for emotional support. AI tools like ChatGPT are preferred for sharing personal thoughts among youth. Concerns regarding social isolation and privacy risks accompany the use of AI for emotional support.
India Today
an hour ago
- India Today
Meta splits AI division into 4 teams, job cuts loom after Mark Zuckerberg's big hiring spree
Meta is once again shuffling its artificial intelligence division, this time splitting it into four separate groups. The move, announced internally this week, is being seen as an attempt to bring order to the company's high-stakes race for advanced AI. But alongside the reorganisation, questions are being raised about possible job cuts, a striking change given how aggressively CEO Mark Zuckerberg has been hiring AI talent over the past to Bloomberg, the internal memo was shared by Alexandr Wang, Meta's newly appointed chief AI officer and former Scale AI CEO. In the note, Wang said the company needed sharper focus to work towards its long-term vision. 'Superintelligence is coming, and in order to take it seriously, we need to organise around the key areas that will be critical to reach it — research, product and infra,' he division, called Meta Superintelligence Labs (MSL), will now function as four distinct teams. One group will look after large language models such as Llama, which powers Meta's AI assistant. Another will continue long-term research under the FAIR banner, a lab that has been part of Meta for over a decade. A third, led by former GitHub CEO Nat Friedman, will focus on bringing AI models into consumer products. The fourth will handle infrastructure like data centres and computing hardware, under the leadership of Aparna Ramani. This restructuring follows months of heavy investment in AI. Bloomberg reported that Meta has been poaching leading researchers from rivals like OpenAI and Google, offering pay packages worth hundreds of millions of dollars. Zuckerberg has also made it clear that he is ready to spend 'hundreds of billions' over the coming years on both talent and the infrastructure required to build what he calls 'superintelligence' — AI that can outperform humans in complex the big hiring spree could soon give way to downsizing. The New York Times reported that Meta is considering cutting or reassigning roles within the AI division, which has grown into a workforce of several thousand. Some executives are also expected to move on, adding to the churn. Although no final decision has been made, the possibility of layoffs has unsettled employees, especially since the company had positioned AI as its biggest area of Meta, this is yet another restructuring in a series of shake-ups around its AI plans. Earlier this year, the AGI foundations group — which worked on generative AI research — was dissolved, and its leaders were reassigned. Connor Hayes, who previously ran AI products, was moved to oversee Threads, Meta's answer to X (formerly Twitter). And Loredana Crisan, who shifted from Messenger to the generative AI team, has left the company to join design platform constant changes highlight both the urgency and the turbulence inside Meta as it tries to keep up with rivals. Companies like OpenAI and Google have already put advanced consumer-facing models in the market, while Meta is still refining its own systems. The company is also considering a change in approach. The New York Times noted that Meta is now exploring whether to use third-party AI models, including open-source and licensed platforms, instead of relying entirely on its own now, the official line is that the split will bring clarity and focus. But with the division's future size still under discussion and top-level exits continuing, the path ahead is far from smooth.- Ends
