ChatGPT Gets Chattier With New Voice Mode

When ChatGPT hyped its new super-smart GPT-4o AI back in May, its technical details were upstaged by the smart new chatbot's extraordinarily human-like voice. Though that turned out to be just a little too similar to that of actress Scarlett Johansson, causing its own mini digital drama, prompting OpenAI to delay rolling out the new "advanced voice mode" system. After some safety upgrades and a less familiar sounding voice, the system is ready for public release to all paying OpenAI users.

For businesses that are already using AI, it may make office work chattier. When you're speaking to ChatGPT, it will be a much more natural-sounding conversation, with the AI using voices and vocal tics that make it seem much more human.

That may change how some users interact with the AI—shedding some of the slightly odd "talking to a robot in my phone" feeling that the system conveyed earlier. Someone trying to impress an AI-hesitant boss could stage a quick demo on the ease and smoothness to win them over. The new voice mode is also accompanied by a slightly refreshed design for the ChatGPT app. TechCrunch reports that some under-the-hood improvements mean that the system is better at understanding spoken commands from people who speak English with a non-native accent.

Time will tell if the super-chatty ChatGPT creeps users out or makes shy people more comfortable engaging with the tech. But it's worth remembering that though the new update makes the AI seem even smarter, this fast-evolving tech doesn't come without risks—including the fact that hackers may see AI as the next digital frontier to cross in an effort to steal your precious data.

Tech news site Ars Technica just reported on a particularly dastardly-sounding new hack which targeted a specific cyber vulnerability in ChatGPT, allowing a malicious coder to "exfiltrate," i.e., steal, information that users had stored in the chatbot's new "long term memories." This feature allows new chats with the AI to quickly reference previous conversations and personal details already shared with the system, and now it may increase the system's vulnerability.

But your intimate discussions with ChatGPT about how many "R"s are in the word strawberry won't leak out—this loophole was spotted and tested out by a security researcher rather than a hacker. It involves planting "false" memories inside ChatGPT's system that include instructions hidden inside uploaded content like emails, blog posts, or documents. The malicious code can be activated to help hackers, for example, by forwarding every future interaction with the chatbot to the hacker's website. ChatGPT has already put a partial fix in place, Ars reports, but points out that business users of ChatGPT need to be constantly wary of the proprietary info they share with the app's memories.

Meanwhile, even as ChatGPT is stealing all the AI limelight, one of its big tech rivals—Google—clearly felt the need to shine a spotlight on its own, expensive AI systems, like Gemini.

In a new blog post arguing that really, honestly, lots and lots of people really are using Gemini at work, Google says it's been improving all sorts of its AI tech and "all of this innovation is driving incredible use of our products." Does the tech giant protest too much?

The post also showcases how companies like Volkswagen and Snap are using Gemini, and it highlights how practical some of its AI "agents" can be when deployed to help people streamline their office work. That last section is interesting because agents are widely touted as the next big thing in AI—they're little AI software engines that can "do" stuff like filling in forms on your work PC, versus the reactive question-answer system of a typical chatbot.

Google's blog post is full of PR bluster, and its timing relative to the new ChatGPT rollout is amusing. But it's at least a source of useful ideas about how your own company could leverage AI to boost business.

At Spectro Agency, we recognize the incredible potential that AI innovations, like ChatGPT’s new voice mode and Google’s Gemini, bring to businesses. As leaders in high-end digital marketing, app creation, AI-powered solutions, chatbots, software creation, and website development, we empower companies to integrate cutting-edge technologies seamlessly into their operations. Visit us at spectroagency.com to discover how we can help transform your business with the latest advancements in AI.

Source: [Kit Eaton, Inc.com](https://www.inc.com/kit-eaton/chatgpt-gets-chattier-with-new-voice-mode.html)