For prompts that Microsoft’s Prometheus system judges as simpler, Bing chat generates responses using Microsoft’s homegrown Turing language models, which consume less computing power and are more affordable to operate than the bigger and more well-rounded GPT-4 model. Bing’s conversational answers do not always draw on GPT-4, Ribas says. “The implementation in the real world of OpenAI models should be slowed down until all of us, including OpenAI and Microsoft, better study and mitigate the vulnerabilities,” says Jim Dempsey, an internet policy scholar at Stanford University researching AI safety risks.Īt the same time, Microsoft has held back from going all-in on OpenAI’s technology. Critics, including some of Microsoft’s own employees, warn of potential harms such as AI-crafted misinformation, and some have called for a pause in further development of systems like Bing chat. Bing chat’s answers occasionally are misleading or outdated, and the service, like other chatbots, can be annoyingly slow to respond. Some users held court with Bing chat for hours, exploring conversational paths that led to unhinged responses Microsoft responded by instituting usage limits. Since then, millions of people spanning 169 countries have used it for over 100 million conversations. Prometheus became the foundation for Bing’s new chatbot interface, which launched in February. In October, they showed him a prototype of a search tool the company calls Prometheus, which combines the general knowledge and problem-solving abilities of GPT-4 and similar language models with the Microsoft Bing search engine. Ribas again challenged the system in his native languages, posing Prometheus complex problems like vacation planning. Ribas asked some of Microsoft’s brightest minds to probe further. “That's when we had that ‘aha’ moment,” Ribas says. When quizzed about history, churches, and museums, its responses hit the mark. Then he asked GPT-4 to solve an electronics problem about the current flowing through a circuit. The bot nailed it. As Ribas had with GPT-4’s predecessors, the Barcelona native wrote in Spanish and Catalan to test the AI’s knowledge of cities like his hometown and nearby Manresa. That month, the Microsoft search and AI chief got the keys to GPT-4, a then secret version of OpenAI’s text-generation technology that now powers ChatGPT. At the time, the company was also initially reluctant to acknowledge the problem, but did eventually said it would issue a fix.Jordi Ribas hasn’t taken a day off since last September. Last summer, another bug was found to give away the user's precise location thanks to Google's location services. This isn't the first time Chromecast users have their privacy exposed. In an e-mail to TechCrunch, the company did say it's working on a fix for the "deauth" bug, even if it's four years late to take action. Since the attack was carried out, the dedicated page for tracking the attack has been taken down, and YouTube gave one of the hackers' channels a strike, while also taking down one of his videos. The major difference here is that thanks to UPnP, the attackers were able to carry out the attack over the internet, while previous demonstrations were made over the local Wi-Fi network, thus requiring the attacker to be within its range and able to authenticate with it. This isn't a new issue, either, as security firms started exploiting this "deauth" bug all the way back in 2014. The search giant isn't without fault, though, because the more concerning bug is the fact that Chromecast devices allow an unauthenticated device to access them and control video playback at will. This, however, only refers to one of the vulnerabilities that allowed the attack - Universal Plug and Play (UPnP) - which makes local devices publicly accessible. Google responded to the attack saying that it wasn't a fault in Chromecast but rather in the router settings of the home network of users. Additionally, the broadcasted image asks users to subscribe to PewDiePie, which was a popular joke in the YouTube community a few weeks ago, and one that HackerGiraffe had already participated back in September by hacking printers to advertise the YouTuber. This specific attack was fairly harmless, and the hackers simply posted a message, warning users that their device was exposed to people on the internet, but this does mean that those devices can be hacked with much more malicious intents. The attack exploits a bug that allows virtually anyone to make Chromecast devices play any YouTube video they want. Google's Chromecast devices have been hijacked by a team of two hackers, Twitter users and who remotely hacked into thousands of streaming sticks to broadcast their own custom message.
0 Comments
Leave a Reply. |
Details
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |