关 于 亚信安全科技股份有限公司 2024 年限制性股票激励计划 第一个归属期归属条件未成就 暨 部分限制性股票作废事项 的 法 律 意 见 书 汉坤（证）字[2025]第 20138-10-O-2 号 北京市汉坤律师事务所 中国北京市东长安街 1 号东方广场 C1 座 9 层 100738 电话：(86 10) 8525 5500；传真：(86 10) 8525 5511 / 8525 5522 北京 上海 深圳 香港 海口 武汉 新加坡 纽约 www.hankunlaw.com 北京市汉坤律师事务所 法律意见书 北京市汉坤律师事务所 关于亚信安全科技股份有限公司 2024 年限制性股票激励计划 第一个归属期归属条件未成就暨部分限制性股票作废事项的法律意见书 汉坤（证）字[2025]第 20138-10-O-2 号 致：亚信安全科技股份有限公司 北京市汉坤律师事务所（以下简称"本所"）接受亚信安全科技股份有限公司 （以下简称"亚信安全"或"公司"）委托，担任公司 2024 年限制性股票激励计 划（以下简称"本激励计划"）的特聘专项法律顾问，就公司本激励计划第一个归 属期归属条件未成就暨部分限制性股票作废（以 ...

人工智能技术加速渗透千行百业，网络安全威胁亦随之迭代升级。传统单点防御模式因数据孤岛、响应 滞后等问题，难以应对AI驱动的自动化攻击。大模型滥用导致的深度伪造、数据泄露等新型风险频 发。在此背景下，网络安全产业也必须从"孤立防护"向"体系化联动"转变。 日前，在C3安全大会上，中国信息通信研究院副院长魏亮指出：发展AI+安全，安全应用是核心，既要 关注对不同安全能力的关联整合，也能以攻促防提升安全检测能力。 在大模型引发范式转变的同时，全球大模型安全事件损失急剧增长，AI攻击变异率更是上升到每24小 时93%的惊人水平，导致传统检测系统完全失效，企业安全防御面临前所未有的挑战。 在这种背景下，网络安全需要关注三个核心方向，一是安全内生化，将安全从外部防护转变为系统内在 属性；二是，安全即能力，将安全视为AI系统的核心能力而非限制；三是，生态共建，通过开放合作 构建智能新纪元的安全基础设施。 吴湘宁在采访中强调："工具堆砌反而会导致防御效率下降，在诸多设备中寻找线索、挖掘风险，这个 过程需要的时间太长，而真正的攻击到来时，根本来不及反应。"他提到，传统XDR更多停留在数据采 集层面，而真正的威胁响应需要"智能统 ...

Core Insights - The integration of AI into cybersecurity is creating more complex and frequent security threats, with global losses from large model security incidents projected to rise from $8.5 billion in 2023 to $14.3 billion in 2024, marking an increase of nearly 70% [2] - AI-assisted attacks are evolving rapidly, with a mutation rate of 93% every 24 hours, indicating a significant challenge for traditional security measures [2] Group 1: AI XDR System - The company proposes the AI XDR (Extended Detection and Response) system as a comprehensive solution to counter AI-driven threats, emphasizing a shift from traditional single-point defenses to a systematic, integrated approach [3] - The AI XDR system is described as a suite or scenario-based solution that integrates threat intelligence, behavior analysis, and automated response capabilities into a cohesive security framework [3] - The system aims to deliver a unified interface and a holistic security solution, moving away from the traditional model of delivering individual components [3][5] Group 2: Implementation and Benefits - The AI XDR system allows for a phased implementation, enabling companies to pilot the system in parts of their operations before full-scale deployment [5] - The system significantly reduces alert volumes and minimizes false positives and negatives, thereby alleviating the pressure on security operations personnel [5] - The transition to a new security paradigm is emphasized, where data-driven insights replace experience-based approaches, allowing for better threat perception and decision-making through AI [5]

Core Insights - The integration of artificial intelligence (AI) is transforming the cybersecurity industry, presenting both significant opportunities and challenges for companies to establish robust security defense systems in the "AI+" era [1][3] - The rapid increase in security incidents related to large models is alarming, with losses projected to rise from $8.5 billion in 2023 to $14.3 billion in 2024, and expected to exceed $23.5 billion by 2025 [3] - The mutation rate of AI-assisted attacks is extraordinarily high, reaching 93% every 24 hours in 2024, indicating a shift in security challenges from mere technical issues to critical concerns affecting business models and societal stability [3] Industry Trends - The C3 Security Conference 2025 highlighted the need for a systematic, intelligent, and integrated defense approach to counter AI-driven threats, emphasizing that traditional isolated security measures are no longer sufficient [3][4] - The launch of the AI XDR system by AsiaInfo Security represents a significant advancement in cybersecurity, focusing on data-driven, AI-native architecture to enhance cross-dimensional security capabilities [4][5] - AsiaInfo Security is investing 30% of its annual R&D budget into core technology development, aiming to build a smart defense system supported by leading threat intelligence and AI security models [4][6] Business Model Evolution - The convergence of products and platforms is revolutionizing the delivery of cybersecurity solutions, leading to new business models akin to established SaaS frameworks [5][6] - AsiaInfo Security's collaboration with China Unicom aims to establish a leading domestic SaaS security service brand, reflecting a strategic shift towards integrated service offerings [6] - The introduction of the "Security + Intelligence" integrated capability system during the conference aims to address various security scenarios in the digital economy, showcasing the company's commitment to enhancing cybersecurity in emerging technologies [6][7]

亚信安全科技股份有限公司（以下简称"公司"）于近日收到政府补助人民 币1,647,400.00元，属于与收益相关的政府补助。 二、政府补助的类型及对公司的影响 证券代码：688225 证券简称：亚信安全 公告编号：2025-033 亚信安全科技股份有限公司 关于获得政府补助的公告 本公司董事会及全体董事保证本公告内容不存在任何虚假记载、误导性陈述 或者重大遗漏，并对其内容的真实性、准确性和完整性依法承担法律责任。 一、获得补助的基本情况 公司根据《企业会计准则第16号——政府补助》等有关规定，确认上述事项 并划分补助类型。上述政府补助未经审计，具体的会计处理以及对公司损益的影 响情况最终以审计机构年度审计确认后的结果为准，敬请广大投资者注意投资风 险。 特此公告。 亚信安全科技股份有限公司董事会 2025年5月27日 ...

Group 1 - The C3 Security Conference held in Nanjing on May 24, 2025, focused on the theme "Security + Intelligence, AI Wins the Future," attracting over 2,000 participants from various industries to discuss new paradigms in digital security [1][2] - The conference featured a main forum, five technical and industry sub-forums, a special event called "C3 Technology Night," and an exhibition area showcasing innovative technological achievements [1] - Notable speakers included industry leaders such as Wei Liang, Deputy Director of the China Academy of Information and Communications Technology, and Liu Xiangyang, Chief Information Security Officer of Midea Group, who discussed challenges in AI technology implementation and the evolution of cybersecurity threats [2] Group 2 - A new generation of systematic cybersecurity products, the AI XDR Linked Defense System, was launched by AsiaInfo Security, emphasizing integrated delivery to address cybersecurity threats [3] - AsiaInfo Security and China Unicom jointly released the "Unified Security Detection and Response Service Platform," along with the TrustOne Harmony version application and the "New Generation Terminal Security Harmony White Paper" [3] - AsiaInfo Security will serve as the exclusive cybersecurity partner for the 12th World Games in Chengdu, set to open on August 7, 2023, and announced the "2025 Cybersecurity Assurance System for the 12th World Games" [3]

Core Insights - The C3 Security Conference highlighted the escalating security risks associated with AI, with global losses from large model security incidents projected to rise from $8.5 billion in 2023 to over $23.5 billion by 2025, indicating a significant increase in AI-related security threats [4][6][10] - AI is exacerbating the asymmetry in cybersecurity, with attackers leveraging open-source tools and AI to enhance their attack methods, while defenders struggle with outdated techniques and resource constraints [5][9] Group 1: AI Security Challenges - The rapid mutation rate of AI-assisted attacks is alarming, with a reported 93% mutation rate every 24 hours in 2024 [4] - The attack efficiency of AI surpasses defensive responses, creating a technological imbalance where attackers can exploit publicly available data for training, while defenders face compliance issues [5][10] - The average time for attackers to breach systems is now 48 minutes, with the fastest recorded at 51 seconds, highlighting the urgency of the threat landscape [9] Group 2: Talent Shortage and Industry Response - There is a significant talent gap in the AI security field, with a shortage of over 300,000 professionals globally, particularly those with expertise in both AI and cybersecurity [10] - Companies are adopting a T-shaped talent development model to address the shortage, emphasizing the need for professionals who possess deep security knowledge and broad AI understanding [10] - The industry consensus is shifting towards a systematic and integrated defense approach, moving away from isolated defenses to a collaborative security matrix [11][12] Group 3: Future of Cybersecurity - The next decade is seen as a golden period for cybersecurity, with a predicted evolution towards cognitive security paradigms and self-supervised security systems by 2030 [13] - The integration of AI into security frameworks is expected to blur the lines between security and intelligence, leading to a new paradigm of "security intelligence" [13]

Market Performance - The Zhipu AI concept index declined by 3.09%, ranking among the top losers in the concept sector as of May 15 [1] - Within the Zhipu AI sector, stocks such as Aofei Data, Pingzhi Information, and Hongxin Electronics experienced significant declines, while stocks like Tuosida, Guangyun Technology, and Shengtong Co. saw gains of 4.66%, 2.65%, and 0.97% respectively [1] Capital Flow - The Zhipu AI concept sector saw a net outflow of 3.465 billion yuan, with 55 stocks experiencing net outflows and 12 stocks exceeding 100 million yuan in outflows [2] - The stock with the highest net outflow was Changshan Beiming, with a net outflow of 439 million yuan, followed by Aofei Data, Hand Information, and Tianyuan Dike with net outflows of 361 million yuan, 229 million yuan, and 198 million yuan respectively [2] Top Gainers and Losers - Among the top gainers in the Zhipu AI concept, Guangyun Technology led with a net inflow of 36.43 million yuan, followed by Yinsai Group and Mengwang Technology with inflows of 25.39 million yuan and 22.21 million yuan respectively [4] - Conversely, stocks like Aofei Data and Hongxin Electronics faced significant losses, with Aofei Data down by 7.93% and Hongxin Electronics down by 7.24% [3]

一、获得补助的基本情况 亚信安全科技股份有限公司（以下简称"公司"）于近日收到政府补助人民 币2,709,875.85元，属于与收益相关的政府补助。 二、政府补助的类型及对公司的影响 公司根据《企业会计准则第16号——政府补助》等有关规定，确认上述事项 并划分补助类型。上述政府补助未经审计，具体的会计处理以及对公司损益的影 响情况最终以审计机构年度审计确认后的结果为准，敬请广大投资者注意投资风 险。 特此公告。 亚信安全科技股份有限公司董事会 2025年5月10日 证券代码：688225 证券简称：亚信安全 公告编号：2025-032 亚信安全科技股份有限公司 关于获得政府补助的公告 本公司董事会及全体董事保证本公告内容不存在任何虚假记载、误导性陈述 或者重大遗漏，并对其内容的真实性、准确性和完整性依法承担法律责任。 ...

亚信安全公告，公司于近日收到政府补助270.99万元，属于与收益相关的政府补助。该政府补助未经审 计，具体的会计处理以及对公司损益的影响情况最终以审计机构年度审计确认后的结果为准。 ...