Google doubles down on ‘AI phones’ with its Pixel 10 series
来源：TechCrunch | 发布时间：2025-08-31
摘要：2025年8月20日，Google发布了Pixel 10系列，包括Pixel 10、Pro、Pro XL和Pro Fold，搭载Tensor G5芯片，增强AI功能和摄像头，售价从799美元到1800美元不等，将于8月底上市。Pixel 10系列还推出了Pixelsnap配件。 Google宣布为Google Photos推出新的AI驱动的照片编辑功能，允许使用文本或语音提示。该功能在美国的Pixel 10上可用，包括C2PA内容凭证，使用Gemini AI，用户可以使用“删除汽车”或“恢复旧照片”等命令编辑照片。
With the launch of the new Pixel 10 series, Google is rushing ahead of Apple to deliver AI-powered smartphones to consumers. The devices, announced during Wednesday’s Made by Google event, come just weeks ahead of Apple’s expected iPhone 17 reveal, which promises to be more of the same — better cameras, possibly thinner devices, and new colors to choose from.
Google, meanwhile, has been rapidly integrating its AI platform into its devices.
Last year, its Pixel 9 series added a number of AI features, like Gemini Live(Gemini’s voice mode), image-generation tools, call notes, searchable screenshots, and more. Since then, Google says that Gemini Live conversations have proven to be 5x longer than text-based conversations.
This year, the tech giant is rolling out even more AI-powered upgrades with the launch of its Pixel 10, including a Visual Overlays feature for the camera, a proactive “Magic Cue” feature, Camera Coach, Voice Translate for calls, an assistant-like “Take a Message” feature, Pixel Journal, and more.
Combined, the updates allow Google to showcase what its latest AI technology can do when enhanced by its Tensor G5 processor, an upgrade to the company’s custom silicon designed for AI experiences and the first to run its newest Gemini Nano model.
Alongside the launch, Google announced that Gemini Live will gain a new audio model that will detect your tone — like whether you’re excited or concerned — and adjust its response accordingly.
Another new feature, Magic Cue, lets the AI be more proactive by offering contextual suggestions in real time, across apps like Gmail, Calendar, Messages, Screenshots, and others.
The idea of a more proactive interaction between people and Google technology is something the company has dreamed of for years, long before the AI era. In the early 2010s, for example, Google introduced an Android feature called Google Now that would pop up cards with real-time information related to your daily schedule or the time of day, like nearby restaurants at lunchtime, upcoming meetings, or flight details.
Years later, Magic Cue is the AI-powered reintroduction of this feature, but one where it inserts itself into your everyday apps and interactions.
Google demonstrated how Magic Cue could suggest a restaurant to dine at with a friend, offering quick access to place a call to the restaurant to make a reservation. It could propose a reply to your friend with the reservation details or point you to your calendar to check your availability.
Magic Cue’s suggestions appear within the app you’re using and are wrapped with a rainbow-colored outline to differentiate them, as well as within Daily Hub, a personalized daily digest in your Discover feed. You can also tap on its suggestions to take action.
Also similar to Google Now, Magic Cue will be able to surface reminders. But it goes a step further by popping up reminders and notifications more intuitively. For instance, it may remind you of errands you need to handle, like a return of an online order, suggest topics you may want to research, or recommend new playlists to stream.
At launch, Magic Cue’s suggestions will be limited to select activities, like settling up a tab, adding events to your calendar, and showing the forecast for an upcoming trip in the weather app. Over time, Google will add other options and let you configure which data source the feature has access to.
Apple, it should be noted, is trying to do something similar by allowing users to speak to Siri to interact with and take action within their apps, but unfortunately, its AI-powered Siri has been delayed until 2026.
Another one of the more interesting additions in the Pixel 10 series is Camera Coach, an AI-powered assistant that aims to make you a better photographer.
The feature will be launching in preview with the new devices and uses Gemini models to offer suggestions about how to better frame and compose your shot. You can even choose a “get inspired” option that will suggest scenes you may not have considered.
Plus, the camera will now be able to recognize when you’re taking a group photo. The “Auto Best Take” feature activates and analyzes up to 150 images shot over several seconds to find the best one — whether that’s a shot you snapped yourself or one made by blending others together via AI.
Then, with the AI-powered “Ask Photos” tool, you can edit the shot to do other things, like fix the lighting, change the framing, or remove an object from the photo by either speaking to or texting Photos’ AI assistant.
On Pro devices, the Pro Res Zoom option will also use AI to allow you to “zoom” in on things like architecture and landscapes at 30x-60x or 30x-60x for animals and wildlife.
Related to its enhanced use of AI in photography, Pixel 10 phones will also be the first to implement C2PA, a standard that establishes the origin and edits of digital content, which will help to identify when photos have been modified by AI.
Another new AI feature, Voice Translate, will use on-device AI to translate your phone call in real time in what sounds like each speaker’s own voice. This could be a potential game changer, particularly for business users and world travelers, if it works as well as described. (This still needs to be tested by reviewers, of course.)
The feature will translate to or from English and Spanish, German, Japanese, French, Hindi, Italian, Portuguese, Swedish, Russian, and Indonesian.
In Pixel’s Phone app, a new addition called Take a Message provides real-time transcripts for missed and declined calls and then uses AI to identify the next steps you need to take based on the caller’s voicemail. (That update will come to Call Notes, too.)
The feature will translate to or from English and Spanish, German, Japanese, French, Hindi, Italian, Portuguese, Swedish, Russian, and Indonesian.
In Pixel’s Phone app, a new addition called Take a Message provides real-time transcripts for missed and declined calls and then uses AI to identify the next steps you need to take based on the caller’s voicemail. (That update will come to Call Notes, too.)
Pixel Journal, meanwhile, is Google’s answer to Apple’s Journal app, but one that uses AI to prompt you to share your thoughts, track your progress toward goals, and offer insights over time.
Other minor AI upgrades to the Pixel 10 lineup include writing tools integrated into the Gboard keyboard, updates to Pixel screenshots in Pixel Studio, and Notebook LM integrations with Recorder and screenshots.
谷歌发布Pixel 10系列主打AI手机功能
谷歌发布Pixel 10系列AI手机，抢在苹果iPhone 17之前推出。新设备搭载Tensor G5处理器，配备Visual Overlays相机功能、主动式Magic Cue助手、Camera Coach拍照指导、实时语音翻译等AI功能。Gemini Live新增音调检测，可根据用户情绪调整回应。设备还支持C2PA标准识别AI修图，Voice Translate实现实时通话翻译。
在新推出的Pixel 10系列发布中，谷歌抢在苹果之前为消费者带来AI驱动的智能手机。这些设备在周三的”Made by Google”直播活动中正式发布，比苹果预计的iPhone 17发布提前数周，而iPhone 17预计只会带来常规升级——更好的摄像头、可能更轻薄的设计和新配色选择。
与此同时，谷歌一直在快速将其AI平台整合到设备中。
去年，Pixel 9系列增加了多项AI功能，如Gemini Live（Gemini语音模式）、图像生成工具、通话记录、可搜索截图等。谷歌表示，自那时起，Gemini Live语音对话的时长比文本对话长五倍。
今年，这家科技巨头在Pixel 10发布中推出了更多AI驱动的升级功能，包括相机视觉叠加功能、主动式”Magic Cue”功能、Camera Coach、通话语音翻译、类似助手的”Take a Message”功能、Pixel Journal等。
这些更新的组合让谷歌能够展示其最新AI技术在Tensor G5处理器增强下的能力，这是该公司为AI体验设计的定制芯片的升级版，也是首个运行最新Gemini Nano模型的处理器。
在发布的同时，谷歌宣布Gemini Live将获得新的音频模型，能够检测你的语调——比如你是兴奋还是担忧——并相应调整其回应。
通过名为视觉叠加的新功能，Gemini Live将能够通过相机镜头看到你所看到的内容，并通过在屏幕上高亮显示内容来提供指导。例如，在国外旅行时，你可以举起手机查看周围的街道标志是否提供路边停车信息。
另一个新功能Magic Cue让AI更加主动，在Gmail、日历、信息、截图等应用中实时提供情境建议。
让AI与谷歌技术进行更主动交互的想法是该公司多年来的梦想，早在AI时代之前就有了。例如，在2010年代初，谷歌推出了一个名为Google Now的Android功能，会弹出与你日常安排或时间相关的实时信息卡片，如午餐时间附近的餐厅、即将到来的会议或航班详情。
多年后，Magic Cue是这一功能的AI驱动重新推出，但它会深入到你的日常应用和交互中。
谷歌演示了Magic Cue如何建议与朋友用餐的餐厅，提供快速致电餐厅预订的功能。它可以建议回复朋友预订详情，或指向你的日历检查可用时间。
Magic Cue的建议会出现在你正在使用的应用中，并用彩虹色边框包围以区分，同时也会出现在Daily Hub中，这是你Discover信息流中的个性化每日摘要。你也可以点击建议来采取行动。
与Google Now类似，Magic Cue也能够显示提醒。但它更进一步，能够更直观地弹出提醒和通知。例如，它可能提醒你需要处理的事务，如退回在线订单，建议你可能想要研究的话题，或推荐新的播放列表。
发布时，Magic Cue的建议将限于特定活动，如结算账单、向日历添加事件、在天气应用中显示即将到来的旅行天气预报。随着时间推移，谷歌将添加其他选项，并让你配置该功能可访问的数据源。
值得注意的是，苹果也在尝试类似功能，允许用户通过Siri在应用内进行交互和操作，但不幸的是，其AI驱动的Siri已被推迟到2026年。
Pixel 10系列中另一个有趣的新增功能是Camera Coach，这是一个AI驱动的助手，旨在让你成为更好的摄影师。
该功能将在新设备上以预览版形式推出，使用Gemini模型提供如何更好地构图和拍摄的建议。你甚至可以选择”获得灵感”选项，该选项会建议你可能没有考虑过的场景。
此外，相机现在能够识别你正在拍摄合影，”Auto Best Take”功能会激活并分析几秒钟内拍摄的多达150张图像，找到最佳的一张——无论是你自己拍摄的照片，还是通过AI混合其他照片创建的。
然后，通过AI驱动的”Ask Photos”工具，你可以编辑照片做其他事情，如修正光线、改变构图，或通过语音或文本与Photos的AI助手交流来从照片中移除物体。
在Pro设备上，Pro Res Zoom选项也将使用AI让你能够以30-60倍的倍数”放大”建筑和风景，或以30-60倍放大动物和野生动物。
与其在摄影中增强使用AI相关，Pixel 10手机也将是首批实施C2PA标准的设备，该标准确立数字内容的来源和编辑历史，有助于识别照片何时被AI修改。
另一个新的AI功能——语音翻译，将使用设备端AI实时翻译你的电话通话，听起来像是每个说话者自己的声音。如果效果如描述的那样好，这对商业用户和世界旅行者来说可能是一个潜在的游戏规则改变者。
该功能将支持英语与西班牙语、德语、日语、法语、印地语、意大利语、葡萄牙语、瑞典语、俄语和印尼语之间的翻译。
在Pixel的电话应用中，名为Take a Message的新功能为未接和拒接电话提供实时转录，然后使用AI根据来电者的语音邮件识别你需要采取的下一步行动。
与此同时，Pixel Journal是谷歌对苹果Journal应用的回应，但使用AI提示你分享想法、跟踪目标进展，并随时间提供洞察。
Pixel 10系列的其他小型AI升级包括集成到Gboard键盘的写作工具、Pixel Studio中Pixel截图的更新，以及Notebook LM与录音机和截图的集成。
Q&A
Q1：Pixel 10系列有哪些主要的AI新功能？
A：Pixel 10系列新增了多项AI功能，包括相机视觉叠加功能、主动式Magic Cue功能、Camera Coach摄影助手、通话语音翻译、Take a Message语音留言处理、Pixel Journal日志应用等。这些功能都基于最新的Tensor G5处理器和Gemini Nano模型。
Q2：Magic Cue功能具体能做什么？
A：Magic Cue是一个主动式AI助手，能在Gmail、日历、信息、截图等应用中实时提供情境建议。比如建议与朋友用餐的餐厅并快速预订，建议回复朋友预订详情，提醒你需要处理的事务如退换商品，或推荐研究话题和播放列表等。
Q3：语音翻译功能支持哪些语言？
A：语音翻译功能使用设备端AI实时翻译电话通话，听起来像每个说话者自己的声音。支持英语与西班牙语、德语、日语、法语、印地语、意大利语、葡萄牙语、瑞典语、俄语和印尼语之间的翻译。