计算机 公 2024年04月23日 司 报 科大讯飞(002230.SZ) 告 模型能力持续升级，讯飞星火赋能公司业务发展成果显著 推荐 ( 维持 ) 事项： 公司公告2023年年报。2023年，公司实现营业收入196.50亿元，同比增长 股价:41.97元 4.41%；实现归母净利润6.57亿元，同比增长17.12%。2023年利润分配预案 公 主要数据 为：拟每10股派息1元（含税）。公司同时发布2024年一季报，2024年一季 行业 计算机 司 度，公司实现营业收入36.46亿元，同比增长26.27%；实现归母净亏损3亿 公司网址 www.iflytek.com ...

业绩简评 2024 年 4 月 22日，公司发布 2023 年报及2024 年一季报。公司 2023 年全年实现营收 196.5 亿元，同比增长 4.4%，略低于此前 业绩预告，但达到修订后股权激励营收目标下限。2023 毛利润为 83.8 亿元，同比增长 9.1%；毛利率为 42.7%，较上年增长 1.8pct。 扣除非经常损益的归母净利润为 1.2 亿元，同比下滑 71.7%，接 近业绩预告上限。 分季度而言，公司2023 年第四季度和2024年第一季度分别实现 营收 70.4 亿元、36.5 亿元，分别同比增长14.2%、26.3%；分别 实现扣非后归母净利润4.4 亿元、-4.4 亿元，2023年第四季度扭 亏为盈、2024 年第一季度亏损同比扩大 30.3%。 经营分析 B端及 C端业务增势较高，G 端业务承压。其中2023 年B端智 能汽车网联与智慧金融景气度较高，分别同比增长49.7%、 人民币(元) 成交金额(百万元) 86.00 14,000 23.0%；C端智能硬件受到翻译机同比增长300%等因素的影 78.00 12,000 响，同比增长22.4%；开放平台虽然表观收入同比增长41.6% ...

AI Model Development and Upgrades - The company launched the "iFlytek Spark Cognitive Model" (iFlytek Spark) in 2023, aiming to "liberate productivity and unleash imagination," marking a significant step in the era of general artificial intelligence[2] - iFlytek Spark has been upgraded four times throughout the year, reflecting the company's 24-year commitment to advancing cognitive intelligence, including semantic understanding since 2011 and the iFlytek Brain project since 2014[3] - The company's AI model "Spark" has seven core capabilities including text generation and logical reasoning[11] - The "Xunfei Superbrain 2030 Plan" aims to develop a self-evolving AI system with capabilities surpassing human perception[10] - iFlytek Spark V3.5 model's language understanding, math, and voice interaction capabilities surpass GPT-4 Turbo, with code reaching 96% of GPT-4 Turbo and multimodal understanding reaching 91% of GPT-4V[33] - iFlytek Spark V3.5 model's multilingual speech recognition outperforms OpenAI Whisper V3 in 37 mainstream languages, with an average MOS score improvement of 0.25 in 40 languages and a MOS score of 4.5 for ultra-realistic speech synthesis[33] - The company officially launched the Spark Cognitive Large Model on May 6, 2023, which has become the foundational platform for AI industrialization[31] - The Spark Cognitive Large Model follows a "1+N" layout, with "1" representing the general cognitive intelligence model and "N" representing industry-specific models applied in education, healthcare, automotive, office, industrial, and smart hardware sectors[32] Industry Applications of AI - In the automotive industry, the "Spark Smart Cockpit" has been applied to over 60 overseas models, supporting 23 languages to enhance the competitiveness of Chinese cars in international markets[4] - In the energy sector, iFlytek Spark has been implemented at China National Coal Group, significantly improving procurement efficiency and quality through intelligent solutions[4] - The company's AI programmers have been deployed in over 100 enterprises, including JD.com and SoftStone, enhancing code development efficiency and accelerating digital transformation[4] - In the healthcare sector, the "Smart Medical Assistant" powered by iFlytek Spark has provided over 700 million diagnostic aids across 426 districts in 31 provinces, detecting 54 million unreasonable prescriptions[5] - The company's education products have served over 130 million teachers and students across 32 provinces, promoting educational equity and quality improvement[5] - The company's AI applications span across industries including education, healthcare, and customer service[10] - The company's AI technology includes speech recognition, synthesis, and machine translation[11] - The company's AI technology has been integrated into smart home, smart customer service, mobile terminals, and automotive terminals, among other fields[77] - The company's AI technology has been widely applied in education, healthcare, and urban sectors, with significant market success and profitability[79] - The company's AI products and solutions have been scaled in key national economic sectors such as automotive, finance, and consumer goods[79] Financial Performance - Revenue for 2023 reached 19.65 billion yuan, a 4.41% increase compared to 2022[19] - Net profit attributable to shareholders was 657.31 million yuan, up 17.12% year-over-year[19] - Operating cash flow decreased by 44.55% to 349.76 million yuan in 2023[19] - Total assets grew by 15.13% to 37.83 billion yuan at the end of 2023[19] - Q4 2023 revenue was 7.04 billion yuan, the highest among all quarters[25] - Q4 2023 net profit attributable to shareholders was 557.95 million yuan, significantly higher than previous quarters[25] - The company plans to distribute a cash dividend of RMB 229,317,822.40, with a dividend of RMB 1 per 10 shares, based on a total share capital of 2,293,178,224 shares[7] - Non-current asset disposal gains in 2023 amounted to RMB 22,292,562.80, a significant increase from RMB 4,460,160.01 in 2022 and a reversal from a loss of RMB 7,253,207.50 in 2021[26] - Government subsidies recognized in 2023 totaled RMB 403,935,002.59, a decrease from RMB 473,121,321.03 in 2022 but higher than RMB 438,785,803.74 in 2021[26] - Gains from fair value changes and disposal of financial assets and liabilities in 2023 were RMB 230,342,367.48, a significant improvement from a loss of RMB 216,461,724.78 in 2022 and lower than RMB 338,818,986.25 in 2021[26] - Total non-recurring gains and losses for 2023 were RMB 539,251,197.09, a substantial increase from RMB 143,394,267.19 in 2022 and slightly lower than RMB 577,724,427.75 in 2021[26] Open Platform and Developer Ecosystem - Since the release of iFlytek Spark on May 6, 2023, the iFlytek Open Platform has added 1.676 million developer teams, with the international platform attracting 270,000 overseas developers[6] - The company provides 647 AI capabilities and one-stop AI solutions through its open platform, continuously building the AI industry ecosystem[31] - iFlytek Open Platform added 1.676 million developer teams, a 229% YoY increase, with over 350,000 large model developers and 2.775 million assistant developers creating 49,000 assistant applications[36] - The AI open platform has opened 647 leading AI capabilities and solutions, with over 5.78 million developer teams, a 52% year-on-year increase, and 1.676 million new developers added since the release of the Spark Big Model[57] - The industrial internet platform based on the Spark Big Model saw a 59% increase in total users and an 88% increase in service frequency within a month of its release[57] Education Sector - The company's education products have served over 130 million teachers and students across 32 provinces, promoting educational equity and quality improvement[5] - The "Tailored Teaching" solution has been implemented in 18 regions, including Tianjin, Jiangxi, Tibet, Zhejiang, Shandong, Hunan, and Sichuan, and has been applied in over 55 cities and districts, significantly improving regional teaching quality[39] - The intelligent grading technology has been officially applied in the college entrance examinations of 14 provinces and cities, covering 20 provincial-level college entrance examinations and 108 city-level high school entrance examinations[40] - The "Spark Teacher Assistant" has served over 400 schools, with 90% of teachers reporting improved efficiency in lesson preparation, reducing lesson design time by over 50% and courseware production time by over 60%[41] - The "Spark Cognitive Model" has been integrated into the English subject of the personalized learning handbook, leading to a 40% year-on-year increase in product sales[44] - The AI learning machine has been upgraded to the world's first cognitive model AI learning machine, introducing 8 new features and maintaining the highest Net Promoter Score (NPS) in the industry[44] - The "AI·LH Psychological Partner" has increased the coverage of personalized psychological counseling for students from less than 5% to 52%, reducing the detection rate of depression by 8% and anxiety by 4%[42] - The "Smart Sports" solution has been applied in approximately 1,400 schools in Beijing, Zhejiang, and Anhui, and has been used for intelligent physical education entrance examinations in Hebei, Shandong, and Anhui[43] - The company's AI learning machine T20 series, equipped with the Spark cognitive model, has been developed and is expected to enhance the company's influence in the education sector[101] - The company's AI learning machine T20 series is the first in the industry to receive TÜV Rheinland's natural light display certification, solidifying its leadership in the education tablet market[102] - Developed the Lumie 10 AI learning machine, focusing on "academic + literacy" improvement under the new curriculum standards, enhancing market competitiveness and driving revenue growth[105] - Established partnerships with 45,000 primary and secondary schools nationwide, completing regional base construction and achieving project sales targets[114] - Launched the "Smart Classroom" product, focusing on terminal application value, and effectively improving product market competitiveness and market share[105] - Developed a digital homework system integrated with AI technology, achieving automatic error collection and consolidation exercises, leading competitors[110] - Built a learner-centered academic evaluation system, helping teachers improve teaching precision and students enhance self-learning effectiveness[115] - Completed the development of the after-school service management platform, maintaining the top market share in the sector and expanding brand influence[118] Healthcare Sector - In the healthcare sector, the "Smart Medical Assistant" powered by iFlytek Spark has provided over 700 million diagnostic aids across 426 districts in 31 provinces, detecting 54 million unreasonable prescriptions[5] - The "Smart Medical Assistant" covers over 1,600 diseases and has provided over 740 million AI-assisted diagnostic suggestions, standardized over 280 million electronic medical records, and corrected over 1.3 million diagnostic cases[46] - The "Smart Medical Assistant" has been implemented in over 400 districts and counties across 30 provinces, providing AI-assisted diagnostic suggestions and electronic medical record standardization[46] - The "Spark Medical Model" has passed the national medical health field model application technology standard evaluation, providing AI diagnostic assistance for doctors and personalized post-diagnosis management[45] - The number of diseases that the AI medical assistant can assist in diagnosing has increased from about 1,100 in 2021 to over 1,600 in 2023, with the reliability of initial diagnosis recommendations improving from about 96% to 98% and prescription review reliability increasing from about 93% to 96%[47] - The AI chronic disease management system has been successfully implemented in several regions, including Wuxi Economic Development Zone in Jiangsu, Puyang County in Henan, and Linzi District in Zibo, Shandong[47] - The smart hospital solution at West China Hospital of Sichuan University supports over 13,000 symptoms and 5,000 diseases for pre-diagnosis triage, accurately matching patients to the appropriate departments[48] - The specialized CDSS covers over 50 diseases, and the VTE system has achieved intelligent assessment of over 170 VTE-related evaluation items, with an AI and expert assessment consistency rate exceeding 90%. It serves over 40 of China's top 100 public hospitals, including all top 5 public hospitals[48] - The Anhui Medical Imaging Cloud Platform connects over 1,900 hospitals and has gathered more than 2,700 experts, serving nearly 70 million cases in Anhui with a monthly average online activity of nearly 20,000 cases[50] - The infectious disease monitoring and emergency command information platform in Hefei, Anhui, monitors 41 statutory infectious diseases, 15 non-statutory infectious diseases, and 6 major syndrome surveillance signals, based on data from over 1,400 primary medical institutions and about 20 hospitals[51] - The company has initiated four smart medical insurance pilot projects in Shanxi, Anhui, Jiangsu, and Yunnan, following its first prize in the National Medical Insurance Bureau's Smart Medical Insurance Solution Competition in 2022[52] Smart City and Government Solutions - The smart city business leverages the Spark Cognitive General Model to build industry-specific models and applications, expanding into smart agriculture, smart parks, smart buildings, and smart water conservancy, forming a robust and comprehensive smart city business ecosystem[53] - The digital government business achieved significant breakthroughs, with the "Government Big Model" supporting 22 departments and 30 projects, reducing development cycles by up to 75%[54] - The legal industry big model, compliant with the "Legal Big Model Evaluation Indicators and Methods (Trial Version)", was released, improving case handling efficiency in courts, with judges in Hefei High-Tech Court handling 991 cases per capita, ranking first in the province[55] Automotive Sector - The smart car business saw a record high of over 7.5 million units of pre-installed intelligent products shipped, with a cumulative shipment of 53.6 million units and annual online interactions exceeding 10 billion, a 229% year-on-year increase[74] - The company's FlyFish in-car intelligent audio management system has been applied in over 30 mass-produced models from more than 10 mainstream domestic and joint-venture automakers[73] - The smart cockpit solution has secured platform-level vehicle projects with Chery and Jiangling, with mass production of models equipped with the company's self-developed cockpit domain controller expected in 2024[74] - The company's in-car interaction products support over 20 overseas languages, covering dozens of overseas models from leading automakers such as Chery and SAIC, with mass production in regions including Europe, South America, and Southeast Asia[73] - The company has developed a 100K computing power cockpit domain controller hardware and system based on domestic chips, optimizing the integration of AI algorithms such as voice, vision, and audio effects[73] - The company's intelligent driving solutions achieved their first project win and are expected to enter mass production in 2024, with advanced perception algorithms developed based on the BEV framework[74] Consumer Hardware and AI Products - The smart hearing aid released in 2022 and its upgraded version in 2023 ranked first in sales on JD.com and Tmall during the 2023 "618 Shopping Festival" and has over 55,000 registered users as of January 19, 2024[50] - The "Xunfei Xiaoyi" APP and mini-program, launched in October 2023, significantly improves consultation efficiency and supports standardized medication, reducing potential irrational drug use[50] - Consumer hardware GMV increased by 84% year-on-year, with sales during the 618 and Double 11 shopping festivals growing by 125% and 126% respectively[60] - The company's AI marketing business achieved significant growth during the 618 and Double 11 shopping festivals, with overseas marketing beginning to assist Chinese companies in going global[59] - The company's consumer hardware products gained international recognition at events such as the AI TechDay in Singapore and the Budapest World Athletics Championships[60] - The company's AI marketing business received industry recognition, including the "Best Digital Marketing Platform of the Year" at the Golden Mouse Awards and the "Best Intelligent Marketing Platform of the Year" at the Tiger Roar Awards[59] - iFlytek Smart Office Book series maintains a market share of over 30% and achieved an online comprehensive praise rate of over 97%[61] - The new X3 series of iFlytek Smart Office Book features a 10.65-inch flexible e-ink screen with 300PPI clarity and an eight-microphone array, with sales increasing by over 50% compared to the previous generation[62] - iFlytek Smart Recorder has been the sales champion on JD.com and Tmall for six consecutive years during the Double 11 shopping festival[63] - iFlytek Hearing SaaS service ecosystem has surpassed 100 million users, with paid users increasing by 28.5% year-over-year[64] - iFlytek Hearing's smart briefing system achieved a speech transcription accuracy rate of 96.28% during the 14th National People's Congress[65] - iFlytek Input Method's daily active users increased by 10% year-over-year, with AIGC-driven content value-added revenue growing by 81%[67] - iFlytek Translator's sales increased by over 300% year-over-year, with a market share of over 80% on JD.com and Tmall[68] - iFlytek Translator supported over 200 countries and regions, with sales in overseas markets growing by over 200% year-over-year[69] - iFlytek AI Sweeping Robot X3 Pro received a 100% positive rating on JD.com and won multiple international design awards[70] R&D and Innovation - The company has built a complete and efficient R&D team, which is considered its greatest advantage in the research field[32] - The number of R&D personnel decreased by 4.02% from 9,281 in 2022 to 8,908 in 2023[141] - R&D personnel accounted for 62.05% of the total workforce in 2023, up 0.37% from 61.68% in 2022[141] - R&D investment increased by 14.36% to RMB 3,836,822,735.77 in 2023 from RMB 3,355,146,722.67 in 2022[143] - R&D investment as a percentage of revenue rose to 19.53% in 2023, up 1.70% from 17.83% in 2022[143] - Capitalized R&D investment increased by 13.45% to RMB 1,602,462,883.60 in 2023 from RMB 1,412,519,566.40 in 2022[143] - The AI interactive kit project has been completed, aiming to enhance brand influence and expand the intelligent automotive product matrix[140][149] - The international version of the voice assistant is in the development stage, targeting global automotive users and expanding market size[140] - The intelligent driving assistance system is in the development stage, focusing on L2+ level software and domain controller products[140][142] - The Lumie 10 AI learning machine project has been completed, integrating Spark cognitive model capabilities for educational applications[145][146] - The personalized learning handbook project is in the development stage, leveraging Spark model capabilities for personalized learning solutions[151] Financial and Operational Metrics - The company's core business in software and information technology services accounted for 98.27% of total revenue, with a year-on-year growth of 4.27%[84] - Revenue from the education sector decreased by 9.70% to 5.56 billion yuan, while the teaching business grew by 11.14% to 287.57 million yuan[84] - The open platform business saw significant growth, with revenue increasing by 41.61% to 3.94 billion yuan[84] - Revenue from the smart automotive sector grew by 49.71% to 695.54 million yuan[84] - The company's revenue in the North China region increased by 30.61% to 3.23 billion yuan[84] - The company's gross margin for the software and information technology services sector was 43.10%, with a slight increase of 0.87% year-on-year[88] - The open platform business had a gross margin of 21.35%, a decrease of 6.88% compared to the previous year[89] - Material costs for the software and information technology services industry decreased by 2.66% to RMB

科大讯飞股份有限公司2024年第一季度报告 证券代码：002230 证券简称：科大讯飞 公告编号：2024-027 科大讯飞股份有限公司 2024 年第一季度报告 本公司及董事会全体成员保证信息披露的内容真实、准确、完整，没有虚假记载、误导性陈述或重大遗漏。 重要内容提示： 1.董事会、监事会及董事、监事、高级管理人员保证季度报告的真实、准确、完整，不存在虚假记载、误导性 陈述或重大遗漏，并承担个别和连带的法律责任。 2.公司负责人、主管会计工作负责人及会计机构负责人(会计主管人员)声明：保证季度报告中财务信息的真 实、准确、完整。 3.第一季度报告是否经审计 □是 否 一、主要财务数据 （一） 主要会计数据和财务指标 公司是否需追溯调整或重述以前年度会计数据 □是 否 本报告期 上年同期 本报告期比上年同期增减 营业收入（元） 3,646,215,386.95 2,887,584,461.77 26.27% 归属于上市公司股东的净利润（元） -300,468,030.20 -57,895,269.00 -418.99% 归属于上市公司股东的扣除非经常性 -440,542,522.82 -338,131, ...

文｜《中国企业家》记者 闫俊文 编辑｜李薇 头图来源｜受访者 谁是AI界的余承东，敢说敢干，贴着技术最领先的公司打仗？刘庆峰显然算一个。 20多年前，刘庆峰和科大讯飞的对手是微软、百度等大公司，战场是智能语音；现在，刘庆峰的目光跨越两万公里的太平洋，他只有一个态度：贴着硅谷 一家名为OpenAI的公司打仗。 “GPT-5一出来，我们有可能跟它又有点差距了，但我们还在迎头赶上，我们知道怎么做。”今年全国两会期间，全国人大代表、科大讯飞董事长刘庆峰接 受《中国企业家》独家专访时说。 科大讯飞的作战图是这样的：2023年5月，科大讯飞发布星火认知大模型；6月9日，突破开放式问答，多轮对话能力和数学能力再升级；8月15日，突破代 码能力，多模态交互再升级；10月24日，通用模型对标2022年11月发布的ChatGPT（GPT3.5），中文能力实现超越ChatGPT，英文能力与ChatGPT相当； 到了今年1月，星火认知大模型的数学、语言理解、语音交互能力都超过了GPT-4 Turbo，代码和多模态理解能力也达到了GPT-4 Turbo的九成。 “6个月内，我们可以达到现在全世界最好的GPT-4最优的水平。”刘庆峰说。 ...

投中网 > 商业深度 > 正文 大模型的耐力游戏 光子星球 | 吴坤谚 2024-02-28 10:36:46 豆包仅是字节AI版图中的一隅。 2010年，世界上首款量产纯电车型日产汽车（Nissan）聆风上市的时候，没人能想到日产汽车公司乃至 整个日本汽车行业都将在十余年后的今天沦为陪跑。 类似这样由领跑者沦为追随者的案例并不鲜见，正如在大模型领域，意欲窃取百度、科大讯飞等一众大 模型先行者胜利果实的字节跳动。 众所周知，百度、科大讯飞是国内大模型赛道中先声夺人的代表玩家，早在去年上半年便上线网页测试 版抢占用户心智，以期在正式面向公众服务的时候能快速积累用户。然而据QuestMobile近日发布的数 据，2023年11月，文心一言、豆包与讯飞星火的日活分别是155.4万、124.6万、71.9万。 不难看出，相比于早在2023上半年便上线网页测试的百度与科大讯飞，8月才开启邀测的豆包跑出了更 快的增长速度。 甚至到了2024年2月19日，豆包在iOS免费榜上排名49，而文心一言排名为157，讯飞星火则排名400开 外，类GPT应用的前三把交椅之间已然存在不小的差距。 在今年1月的2024年年度全员会上，字 ...

最近，OpenAI的Sora又在科技圈投入一枚深水炸弹。全球对于大模型的关注，又一次达到高峰。 聚焦到国内，百度、科大讯飞、商汤、华为等大型企业，以及海量的创业小公司都在布局大模型。以往每一次风口吹来的时候，资本总会蜂拥而上。但 是，此次大模型热潮涌来之时，资本的反应很奇怪——讨论度很高但参与度却不高，赛道的投融资并没有想象中火爆。 研究机构CB Insights发布的《2023年人工智能（AI）行业现状报告》（简称“报告”）显示，2023年中国AI领域投融资数量约为232笔，同比下降38%；融 资总额约为20亿美元，同比下降70%。其中，AI大模型的融资活动也没有呈现上升态势。 今年以来，不少投资人对大模型变得更加审慎，“谈得多、投得少”成为常态。某投资人告诉《一点财经》，现在国内资本圈对于大模型持观望态度的机构 变得更多了，因为大家担心大模型在去年经历了一轮爆火之后，变得估值虚高形成了泡沫。 连街头大妈都知道，AI大模型是当下最热门的科技话题，为何在国内大模型似乎受到资本“冷落”？国产大模型，为何没有引爆一场“烧钱游戏”？ 国产大模型，“冷热”分明 上一个像AI大模型一样热闹的风口，还是十年前的移动互联网 ...

引言：当前国内经济正从传统动能驱动转向新动能驱动，新动能是未来增长空间最为广阔的领域。以数 字经济、AI、高端制造为代表的新动能方向近期产业催化不断，有望逐步映射到股市的投资机会中。 我们筛选出2024年产业发展迎来重要变革、有望成为重要投资主线的十大产业趋势，供各位投资者参 考： 一、算力：海外头部加速“内卷”，国产厂商正在崛起 需求端，算力是数字时代的新生产力。数据海量增加，算法模型愈加复杂，应用场景的深入和发展，都 带来了对算力需求的快速提升。AI发展主要依赖两个领域的创新和演进：一是模仿人脑的数学模型和 算法，其次是支持AI训练计算的半导体芯片。AI芯片所代表的算力是人工智能的底层基石，随着大模 型的参数量和训练数据不断增加，芯片对于算力的需求也进一步增长。2018年发布的GPT-1有1.2亿个参 数，而2023年发布的GPT-4的参数量接近1.8万亿。据估计，以训练GPT-4为例，需要约2.5万块算力为 19.5TFLOPS(浮点运算每秒19.5万亿次)的英伟达A100训练近百天才能完成。根据OpenAI数据显示，从 2012年起，全球AI算力的需求增加了超过30万倍。 海外供给端，AI芯片头部厂商 ...

2024 年 02 月 07 日 公司研究 评级：买入 (维持 ) 研究所： [Table_Title] “双提升”方案提振信心，星火大模型将对标 证券分析师： 刘熹 S0350523040001 liux10@ghzq.com.cn GPT4-Turbo ——科大讯飞（ ）点评报告 002230 事件 : 最近一年走势 2024年2月7日，公司制定并发布了“质量回报双提升”行动方案，扎 实提升公司质量和投资价值，彰显了公司对经营质量和市值管理的信心。 投资要点 : “质量回报双提升”方案提振信心，2023年业绩验证经营质量改善 公司实现业绩增长并实施稳定利润分配。自2008年上市至2023年底， 公司营收增长超76倍，市值较发行市值增长超78倍，近三年以现金累 计分配利润占近三年年均可分配利润比例达118%。未来，公司将重视企 相对沪深300表现 2024/02/07 业发展业绩增长、定期制定《未来三年股东回报规划》、强化信息披露 表现 1M 3M 12M 与公司治理等，实现长期、稳定、可持续的投资者回报。 科大讯飞 2.6% -12.7% -9.9% 沪深300 0.4% -7.6% -18.3% 1月 ...

【TechWeb】2月5日消息，天眼查App显示，近日，苏炳添与北京中科大讯飞信息科技有限公司网络侵 权责任纠纷一审判决书公布。 文书显示，原告苏炳添诉称，讯飞公司于2021年8月、12月分别在其官方微信公众号“躺倒鸭”中未经苏 炳添授权许可，擅自使用苏炳添多张肖像图片，并在文章显著位置植入广告宣传语等商业信息，严重侵 害了苏炳添的肖像权。讯飞公司辩称，相关内容均为转载自媒体报道或微博，案涉文章内容与“躺倒 鸭”产品并无关系，没有借用苏炳添的名誉进行商业推广与宣传。 法院审理认为，讯飞公司未经苏炳添许可，在案涉文章中使用了其肖像照片，并宣传了讯飞公司经营的 业务等。案涉推文会因苏炳添的知名度提高讯飞公司的曝光度和知名度，达到商业宣传的目的，为讯飞 公司带来潜在的商业利益，已构成利用网络侵害苏炳添的肖像权。最终，法院判决讯飞公司在案涉微信 公众号上公开致歉，并赔偿苏炳添经济损失3000元。 ...