ChatGPT boss warns against relying on AI as primary source of information

Following the recent launch of its advanced language model, GPT-5, OpenAI is emphasizing the importance of treating this AI technology as a supplementary tool rather than the primary source of information. In an interview with The Verge, Nick Turley, the Head of ChatGPT, highlighted that while GPT-5 shows marked improvements in accuracy, reasoning, and multimodal capabilities, users should remain vigilant about its limitations. Despite OpenAI's claims of significant progress in minimizing errors, Turley acknowledged that the new model still experiences hallucinations—instances where the AI generates incorrect or misleading information—about 10% of the time. He remarked, "The thing, though, with reliability is that there’s a strong discontinuity between very reliable and 100 percent reliable, in terms of the way that you conceive of the product." Turley advised that until AI can consistently outperform human experts across all subjects, users should verify information obtained from ChatGPT. He believes that people will continue to see the AI as a valuable second opinion instead of their main source of facts. The mechanics behind large language models (LLMs) involve predicting subsequent words based on training data patterns. This process can lead to the generation of seemingly convincing yet inaccurate responses, especially when inquiring about topics not covered in the training data. Turley pointed out that ChatGPT achieves its best performance when used alongside credible external sources, such as search engines. He stated, "I still believe that, no question, the right product is LLMs connected to ground truth, and that’s why we brought search to ChatGPT and I think that makes a huge difference." Looking ahead, Turley expressed optimism regarding the eventual resolution of hallucination issues, although he admitted that significant breakthroughs may not be imminent. Additionally, reports indicate that OpenAI is exploring the development of its web browser and has shown interest in acquiring Google’s Chrome browser should it become available for sale. As the landscape of AI continues to evolve, the need for caution and verification remains paramount for users relying on these innovative technologies.