OpenAI has launched updated ChatGPT with live web search capabilities, aiming to rival Google by providing real-time information and citing sources. This new feature is available for Plus and Team subscribers, using partnerships with trusted news sources for greater accuracy. However, concerns have emerged regarding the reliability of AI transcription tools, like OpenAI’s Whisper, which reportedly fabricates content, raising alarms in sectors like healthcare. Additionally, the White House is treating AI technology as a national security asset, requiring approvals for international sharing. Meanwhile, Google is developing an AI project to automate web browsing, hinting at a future where AI could manage complex online tasks independently.
OpenAI just made a big move by launching live web search for ChatGPT, aiming to challenge Google’s hold on the search Market. This new feature lets users get real-time information on topics like weather and stocks while providing sources for the answers. By partnering with reputable news organizations like the Associated Press and Reuters, ChatGPT is set to deliver more reliable information, potentially changing how people search online.
However, not all news in the AI world is positive. Recent studies show that OpenAI’s transcription software, Whisper, has significant accuracy issues. It has been reported to “hallucinate,” meaning it makes up dialogue that was never actually said. This issue is alarming, especially as healthcare providers start using such technology, with research indicating that about 80% of Whisper’s transcripts are inaccurate.
In other developments, a spinoff from MIT called Liquid AI has created incredibly small neural networks that can perform tasks similarly to much larger systems, like guiding self-driving cars. These tiny networks, with just 19 neurons, could save businesses a lot of money in computing costs.
Meanwhile, the Biden administration is classifying artificial intelligence as a national security asset. Tech giants will now need government approval before sharing their AI advancements overseas. This shift could lead to stricter export regulations akin to those for military technology.
Lastly, Google is reportedly developing an AI project, called “Project Jarvis,” which will allow AI to operate web browsers independently. Set to launch with new Google models in December, this AI could change the way we navigate the internet, performing tasks like filling out forms and managing online shopping.
Stay tuned for more updates on these important developments in AI and technology.
Tags: AI, ChatGPT, Google, OpenAI, technology, national security, AI hallucinations, transcription software, Project Jarvis.
-
What is ChatGPT Search?
ChatGPT Search is a new feature that lets users ask questions and get answers directly from the ChatGPT model, making it easier to find information online. -
Why are there safety concerns with ChatGPT Search?
Safety concerns arise because the model can sometimes provide incorrect or harmful information, leading to potential risks for users who might rely on it for accurate guidance. -
How does OpenAI address these safety concerns?
OpenAI works on improving the model by continuously updating it, gathering user feedback, and implementing filters to reduce the chances of harmful content being shared. -
Can I still trust the information from ChatGPT Search?
While ChatGPT Search can provide helpful information, it’s always a good idea to double-check important facts with reliable sources before taking action based on the answers. - Is ChatGPT Search available to everyone?
Yes, ChatGPT Search is being rolled out gradually, so many users will soon have access to this feature, but some may need to wait a bit longer.
