Core Insights - The AI sector is entering a phase of rapid innovation, with advancements such as DeepSeek's integration of algorithm innovation and engineering optimization, and the release of Llama 4 accelerating the deployment of multimodal large models [1] - Privacy and data security are significant challenges faced by the AI large model field, as discussed at the 2025 Lenovo Innovation Technology Conference [1] Group 1: Data Security and Privacy Protection - Lenovo's CEO Yang Yuanqing emphasized the importance of hybrid AI from the perspectives of personal data protection, enterprise data protection, personal privacy protection, and corporate intellectual property protection [3] - Yang stated that data security and privacy protection are foundational to the three main functions of "super intelligent agents," which represent non-negotiable value propositions and innovation principles [3] - The contradiction between the need for accurate responses and the desire to protect personal data is highlighted, suggesting that personal trusted computing can be a solution by utilizing users' own devices for data learning and inference [3] Group 2: Technological Advancements and Applications - Lenovo's CTO Tolga Kurtoglu noted that data security and privacy protection are crucial in developing "super intelligent agents," advocating for local processing of sensitive information on devices [4] - The need for robust data security measures when transferring data to the cloud is emphasized, along with the importance of anonymizing sensitive information [4] - Yang revealed that Lenovo plans to introduce Deepfake detection technology to mitigate the risk of personal image theft, showcasing the company's commitment to enhancing data security [4]