新浪科技讯 12月8日晚间消息，今日，抖音集团副总裁李亮在个人微博表示，以前在影视作品中看到的 场景，从技术上已经可以实现了。 他以手机AI助手打车为例，指出整个过程"机主"可以通过"助手"丝滑完成，但其中需要调用地理位置、 打车软件、地图导航、路况信息、支付等多个APP及数据，本质上是"机主"授权"助手"完成了这一切， 与什么所谓入侵、黑客、隐私都没有关系。 李亮认为，这只是AI数以万计的应用中最普通的场景之一，距离实现这一切的路已经很近了，但又困 难重重。但无论如何，这些终将都会实现。 责任编辑：何俊熹 新浪科技讯 12月8日晚间消息，今日，抖音集团副总裁李亮在个人微博表示，以前在影视作品中看到的 场景，从技术上已经可以实现了。 他以手机AI助手打车为例，指出整个过程"机主"可以通过"助手"丝滑完成，但其中需要调用地理位置、 打车软件、地图导航、路况信息、支付等多个APP及数据，本质上是"机主"授权"助手"完成了这一切， 与什么所谓入侵、黑客、隐私都没有关系。 李亮认为，这只是AI数以万计的应用中最普通的场景之一，距离实现这一切的路已经很近了，但又困 难重重。但无论如何，这些终将都会实现。 责任编辑：何俊 ...

Core Insights - The penetration rate of AI assistants in smartphones is lower than expected, with users still waiting for a truly powerful and reliable AI assistant [1] - The upcoming flagship phone releases have made AI capabilities a central marketing focus for major manufacturers [1] Group 1: Evolution Trends of AI Assistants - **Deeper Understanding**: AI can recognize more screen information, but this raises privacy concerns, prompting regulatory attention [2] - **Stronger Actionability**: AI is beginning to participate in local life by recognizing restaurants, comparing prices across platforms, and writing reviews [3] - **Higher Interoperability**: There is increased communication and collaboration between different AI assistants [4] Group 2: Key Features and Upgrades - **Screen Content Understanding**: AI assistants are expected to interact more with users' screen content, such as identifying unknown callers and providing alerts for sensitive applications [5] - **Memory Functionality**: New features like "AI one-click flash memory" allow users to summarize and save screen content automatically [6] - **Privacy Measures**: OPPO is addressing privacy concerns by processing sensitive data locally on the device [7] Group 3: AI in Offline Life - **Real-time Interaction**: AI assistants can now use cameras for real-time conversations, enhancing user experience in physical environments [8] - **Price Comparison**: New AI features allow users to compare prices across multiple e-commerce platforms in real-time [8] - **Voice Recognition**: AI assistants are incorporating voice recognition to improve usability in noisy environments [8] Group 4: Challenges and Future Directions - **Complex Task Execution**: Current mobile AI assistants struggle with executing complex multi-step tasks due to a lack of mainstream protocols for communication [12] - **Third-party App Integration**: The reliance on "accessibility features" for AI to control third-party apps raises privacy concerns [12] - **Collaborative AI Models**: The emergence of "smart agents" and A2A protocols may facilitate better task delegation among AI systems [12][13]

Core Insights - The penetration rate of AI assistants in smartphones is lower than expected, with users still waiting for a truly powerful and reliable AI assistant [1] - The upcoming flagship smartphone releases have made AI capabilities a central marketing focus for major manufacturers, indicating a competitive landscape [1] Group 1: Evolution of AI Assistants - The first evolution focuses on deeper understanding, where AI can recognize more screen information, but this raises privacy concerns [1] - The second evolution emphasizes stronger action capabilities, allowing AI to participate in local life by recognizing restaurants, comparing prices, and writing reviews [1][10] - The third evolution highlights increased interoperability among different AI assistants, enabling them to communicate and collaborate more effectively [1][15] Group 2: Market Performance - In Q3, vivo led the domestic smartphone market with 11.8 million units shipped, capturing 17.3% market share, followed by Apple and Huawei [2] Group 3: Understanding Screen Content - Personalization has become a key focus, with AI expected to interact more with users' screen content, such as providing answers while browsing social media [4] - New features like call identification have been introduced to detect potential scams during incoming calls [4] - AI memory functions have been added to summarize screen content and generate bills, although privacy concerns remain [5][8] Group 4: Participation in Offline Life - AI assistants are now equipped with real-time visual capabilities, allowing them to interact with the physical world through camera input [10] - Features like cross-platform price comparison and automatic review generation are being developed to enhance user experience [10][11] Group 5: Interoperability of Intelligent Agents - The lack of mainstream protocols for mobile AI assistants limits their ability to perform complex tasks across different applications [15] - New collaboration paradigms are emerging, with many apps developing their own AI assistants and emphasizing agent-to-agent communication [15][16]

Core Insights - The penetration rate of AI assistants in smartphones is lower than expected, with users still waiting for a truly powerful and reliable AI assistant [1] - The upcoming flagship smartphone releases have made AI capabilities a central marketing focus for major manufacturers, indicating a competitive landscape [1] AI Assistant Evolution Trends - Deeper Understanding: AI can recognize more screen information, but this raises privacy concerns, prompting regulatory attention [1] - Enhanced Actionability: AI is beginning to engage in local life activities, such as restaurant recognition and price comparisons across platforms [1] - Increased Interoperability: There is more communication and collaboration between different AI assistants [1] Market Performance - In Q3, vivo led the domestic market with 11.8 million units shipped, capturing 17.3% market share, followed by Apple and Huawei [2] - Personalization has become a key theme among manufacturers, with AI expected to interact more with users' screen content [2] Memory Functionality - AI assistants from vivo and OPPO have introduced memory features to summarize screen content and generate notes, raising discussions about privacy [3] - OPPO's "AI One-Key Flash Memory" allows users to summarize content from apps like Xiaohongshu, but privacy concerns have led to limitations in functionality with certain apps [3] Regulatory Attention - Regulatory bodies are monitoring the use of screenshot recognition and automated operations by AI assistants, as seen in updates from Honor and OPPO [4] User Experience Challenges - Users find the fragmented AI functionalities across different systems inconvenient, with some expressing that they are less effective than dedicated AI applications [5] - The authenticity of AI-generated reviews and comments is under scrutiny, as distinguishing between genuine and fake information becomes increasingly difficult [6] Collaboration Paradigms - The emergence of "smart agents" and A2A protocols is facilitating communication between different AI assistants, although a mainstream collaboration solution is still lacking on mobile platforms [7] - Huawei's HarmonyOS 6 update emphasizes the coordination of multiple smart agents to assist users in various tasks [7]