内容来源：https://www.wired.com/story/3-tricks-google-gemini-live-latest-major-upgrade/

内容总结：

谷歌旗下语音交互AI助手Gemini Live近日迎来上线约一年半以来“规模最大的更新”。此次升级着重提升语音交互的自然度与对话流畅感，在语音语调、节奏韵律及语义理解等方面实现优化。

用户可通过安卓或iOS版Gemini应用点击右下角声波图标进入Live模式。更新后三大亮点功能值得关注：

沉浸式叙事体验
该模式现可为历史解说、儿童睡前故事、创意脑暴等场景注入情感张力与语音变化。例如可要求AI以凯撒大帝视角讲述罗马帝国兴衰，或分角色演绎《傲慢与偏见》中班纳特姐妹的不同叙事视角，还能生成特定地区数百年前的历史生活图景。

个性化学习助手
在教育辅导场景中，用户可自由设定学习节奏与时长。无论是人类遗传学精要、地毯清洁技巧还是语言学习，AI均可按需调整语速、进行重复讲解。谷歌同时提醒，涉及电路改造、汽车维修等专业操作时，仍需交叉验证信息准确性。

多语种口音模拟
新增的语音模拟功能支持用牛仔口吻讲述西部拓荒史，或以纯正伦敦腔解读英国王室秘闻。语言学习者还可模仿母语者的发音韵律。系统已设置防护机制，对涉及歧视性表达或真人模仿的请求将予以拒绝。

此次升级虽未改变界面设计，多数应答形式也保持原样，但在特定场景中能明显感受到对话质感与个性化程度的提升。

中文翻译：

Gemini Live 是一种更接近对话、更自然的语音交互方式，让您能用声音与谷歌Gemini人工智能助手交流。其理念在于，您可以像和朋友聊天那样与它交谈，可以随时插话打断，尽管实际获得的答案与您平时输入文字查询得到的结果并无二致。

在问世约一年半后，Gemini Live迎来了谷歌所称的"有史以来最大更新"。此次更新使Gemini Live模式比以往更自然、更具对话感，能更好地理解语气、细微差别、发音和节奏。

虽然表面上可能看不出明显变化，许多回复听起来也与从前相似，但在某些特定领域，您能真切感受到最新升级带来的不同——以下就是充分发挥新版Gemini Live功能的方法。

目前这项更新正在安卓和iOS版的Gemini应用中逐步推送。要使用Gemini Live，请启动Gemini应用，点击右下角的Live按钮（图标类似声波图案），然后开始说话。

聆听故事篇章

如今Gemini Live能为讲述的故事注入更多情感与变化——这对历史教学、儿童睡前故事和创意头脑风暴都大有裨益。人工智能甚至会在适当时加入不同口音和语调，帮助您区分角色与场景。

谷歌自己举例说明最佳使用方式：让Gemini Live以凯撒大帝的视角讲述罗马帝国历史。这对Gemini而言是个挑战，需要视角转换和想象力飞跃，并恰当运用语气与风格——而这正是新版Gemini Live更擅长的领域。

您不必局限于凯撒大帝或罗马帝国。例如，可以让Gemini Live分别从班纳特家五位姐妹的视角重述《傲慢与偏见》，或让人工智能编织一段关于您所在地区100年、200年或300年前生活图景的故事。

学习实用技能

Gemini Live新功能显著提升的另一个领域是教育与讲解：您可以要求它就任意主题提供速成课程（或详细教程），无论是人类遗传学的复杂原理，还是清洁地毯的最佳方法。您甚至可以让Gemini Live教您一门语言。

现在人工智能可以按适合您的节奏进行讲解，这在学习新知识时尤为实用。如果您需要Gemini Live放慢语速、加快节奏或重复内容，直接告知即可。如果时间有限，也可以在对话时提前说明。

一如既往，请警惕人工智能的幻觉现象，或许不应完全相信听到的所有内容都绝对准确或经过验证。如果您想学习如何改造家中照明线路或修理故障汽车发动机，请务必通过其他渠道核实所得指导，但Gemini Live至少能成为有用的入门指南。

体验多样口音

本次更新为Gemini Live带来的新技能之一是能用不同口音说话。或许您想听牛仔讲述西部拓荒史，或需要纯正伦敦口音者解说英国王室秘辛。Gemini Live现在都能满足这些要求。

这项功能也延伸至上述语言学习领域，因为您可以听到母语者地道的单词和短语发音，进而尝试模仿其语音语调。虽然Gemini Live尚未覆盖全球所有语言和口音，但已能支持相当丰富的种类。

Gemini Live内置了安全防护机制，如果您的要求涉及贬损性口音使用或试图模仿真实人物，请求可能会被拒绝。不过这仍是测试人工智能的有趣方式，能获得更多样化、个性化的回应。

英文来源：

Gemini Live is the more conversational, natural language way of interacting with the Google Gemini AI bot using your voice. The idea is you chat with it like you would chat with a friend, interruptions and all, even if the actual answers are the same as you'd get from typing your queries into Gemini as normal.
Now, about a year and a half after its debut, Gemini Live has been given what Google is describing as its “biggest update ever.” The update makes the Gemini Live mode even more natural and even more conversational than before, with a better understanding of tone, nuance, pronunciation, and rhythm.
There's no real visible indication that anything has changed, and often a lot of the responses will seem the same as before too. However, there are certain areas where you can tell the difference that the latest upgrade has made—and so here's how to make the most of the new and improved Gemini Live.
The update is rolling out now for Gemini on Android and iOS. To access Gemini Live, launch the Gemini app, then tap the Live button in the lower right hand corner (it looks vaguely like a sound wave) and start talking.
Hear Some Stories
Gemini Live can now add more feeling and variation to its storytelling capabilities—which can be useful for history lessons, bedtimes for the children, and creative brainstorming. The AI will even add in different accents and tones where appropriate, to help you distinguish between the characters and scenes.
One of Google's own examples for how this works best is to get Gemini Live to tell you the story of the Roman Empire from the perspective of Julius Caesar. It's a challenge for Gemini that requires some leaps in perspective and imagination, and to use tone and style appropriately in a way that Gemini Live should now be better at.
You don't have to restrict yourself to Julius Caesar or the Roman Empire either. You could get Gemini Live to give you a retelling of Pride and Prejudice from the perspective of each different Bennett sister, for example, or have the AI spin up a tale of what life would have been like in your part of the world 100, 200, or 300 years ago.
Learn Some Skills
Another area where Gemini Live's new capabilities make a noticeable difference is in educating and explaining: You can get it to give you a crash course (or a longer tutorial) on any topic of your choosing, anything from the intricacies of human genetics to the best ways to clean a carpet. You can even get Gemini Live to teach you a language.
The AI can now go at a pace to suit you, which is particularly useful when you're trying to learn something new. If you need Gemini Live to slow down, speed up, or repeat something, then just say so. If you've only got a certain amount of time spare, let Gemini know when you're chatting to it.
As usual, be wary of AI hallucinations, and maybe don't trust that everything you hear is fully accurate or verified. If you're wanting to learn something like how to rewire the lighting in your home or fix a problematic car engine, double-check the guidance you're getting with other sources, but Gemini Live is at least a useful starting point.
Test Some Accents
One of the new skills that Gemini Live has with this latest update is the ability to speak in different accents. Perhaps you want the history of the Wild West spoken by a cowboy, or you need the intricacies of the British Royal Family explained by someone with an authentic London accent. Gemini Live can now handle these requests.
This extends to the language learning mentioned above, because you can hear words and phrases spoken as they would be by native speakers—and then try to copy the pronunciation and phrasing. While Gemini Live doesn't cover every language and accent across the globe, it can access plenty of them.
There are certain safeguards built into Gemini Live here, and your requests might get refused if you veer too close to derogatory uses of accents and speech, or if you're trying to impersonate real people. However, it's another fun way to test out the AI, and to get responses that are more varied and personalized.