BY BEN SHERRY, STAFF REPORTER@BENLUCASSHERRY
ChatGPT will see you now. And talk to you.
Artificial intelligence company OpenAI plans to roll out voice and image capabilities to ChatGPT users who pay for the chatbot's premium membership or its enterprise version. The voice capabilities will allow users to have spoken conversations with the chatbot, similar to how users interact with virtual assistants like Apple's Siri and Amazon's Alexa, while the image capabilities will allow ChatGPT to analyze uploaded images and answer questions about them. Both could lead to new applications for businesses.
Currently, the only way to interact with ChatGPT is through text. With the new vocal functionality, users will be able to choose between five voice options, which OpenAI says were created in collaboration with professional voice actors. In its announcement, the company provided examples of how the voice assistant could be used to workshop speeches, rehearse presentations, and answer general questions like, "where does the phrase 'potato, potahto' come from?"
As for what integrating vision capabilities into ChatGPT will look like, OpenAI says that users can take a photo, upload it to ChatGPT, and ask for analysis. In an example, a ChatGPT user takes a photo of a bike and asks the chatbot for help lowering the seat. The chatbot provides a method for lowering the seat, then says "if you have tools, show me and I'll guide you further." OpenAI says you could even take a photo of your fridge and ask the chatbot to suggest a meal that could be assembled with the visible ingredients.
To illustrate how ChatGPT's new vision capabilities could be used by businesses, the company simultaneously announced that it had helped develop an A.I. assistant for the Danish company Be My Eyes, which produces a free mobile app that connects seeing volunteers with blind and low-vision people via video conferencing, in order to help them with everyday tasks. In a press release, Be My Eyes announced that they had collaborated with OpenAI to create Be My AI, a new function in the app that allows users to snap a photo of the world around them, and receive an A.I.-generated description. Be My Eyes says that the A.I. app "is perfect for all those circumstances when you want a quick solution or you don't feel like talking to another person to get visual assistance."
The news of OpenAI's new tools came just days before Facebook parent Meta's annual Connect conference, where founder Mark Zuckerberg is expected to reveal his own line of virtual assistants, the Financial Times reported.
Despite OpenAI's new functionalities for ChatGPT, Mor Naaman, a professor of informational science at Cornell University, believes companies need to exercise caution when using the upgraded ChatGPT at work. Business leaders "should be worried about their workers mistaking the fluency of ChatGPT for expertise," Naaman says. "The exchanges with these models may feel satisfying, but we know they are still not reliable and trustworthy enough to use in different contexts without human expertise and evaluation."
Comments