What’s the Point of Chatting with Gemini Live?
Have you ever chatted with a bot that seems unreliable and lacks personality? That’s the question that has been on my mind as I tested Gemini Live, Google’s latest take on OpenAI’s Advanced Voice Mode, last week. Gemini Live aims to provide a more engaging chatbot experience with realistic voices and the ability to interrupt the bot at any time.
Gemini Live has been designed to offer intuitive and interactive conversations, according to Sissie Hsiao, GM for Gemini experiences at Google. It is meant to provide information more succinctly and engage in a more conversational manner compared to traditional text interactions. The goal is to create an AI assistant that can solve complex problems while feeling natural and fluid during interactions.
While Gemini Live does offer a more free-flowing and natural experience compared to Google’s previous AI voice efforts, it still falls short in certain areas. The underlying technology still has issues such as hallucinations and inconsistencies, and Gemini Live introduces some new challenges as well.
The Un-Uncanny Valley
Gemini Live essentially combines text-to-speech technology with Google’s latest generative AI models. The result is a chatbot experience with realistic voices, including options like Ursa, a “mid-range” and “engaged” voice. While the voices are an improvement in expressiveness over older Google voices, they still lack certain qualities like laughter or natural hesitations that can make conversations more engaging.
Unfortunately, Gemini Live tends to maintain a dispassionate tone, making it feel like a polite but unemotional assistant. This can detract from the overall conversational experience, especially compared to more expressive AI models like Advanced Voice Mode.
Chatting with Ursa
During my interactions with Gemini Live, I explored various scenarios, including using the chatbot for job interview prep. While Gemini Live offered some helpful feedback, there were instances where it provided inaccurate information or made up responses. This lack of reliability can make it difficult to trust the chatbot’s advice.
In addition, Gemini Live sometimes struggles with recalling specific details and responding to queries about current events or controversial topics. The responses can be generic and nonspecific, limiting the usefulness of the chatbot in certain contexts.
In Search of Purpose
Despite its promising features, Gemini Live has several technical issues that can impact the overall user experience. From voice cut-outs to response recognition problems, using Gemini Live can sometimes be frustrating. Additionally, the lack of support for certain integrations that are available in the text-based Gemini chatbot further limits its functionality.
Overall, while Gemini Live shows potential, it still feels like a work in progress. The current version may not offer enough advantages over the text-based Gemini experience to justify its exclusive availability on Google’s premium plan. With future updates promising image and real-time video interpretation, Gemini Live may evolve into a more versatile and reliable chatbot.
As the chatbot itself pointed out, there is room for improvement in the interactions and functionality of Gemini Live. Despite its shortcomings, the chatbot provides an interesting glimpse into the possibilities of AI-powered conversations, with both challenges and opportunities for growth.
