正文：
当地时间周四（8月1日），欧盟《人工智能法案》正式生效，这是全球首部全面监管人工智能的法规，标志着欧盟在规范快速发展的人工智能应用方面迈出了重要一步。该法案的实施将分阶段进行，旨在为企业和系统提供必要的过渡期，以便进行相应的调整和修改。
欧盟《人工智能法案》采用基于风险的方法来监管人工智能，这意味着不同应用将根据其对社会构成的风险程度接受不同的监管。该法案将从2025年2月开始实施一系列规则，禁止人工智能应用程序利用个人漏洞，禁止非针对性地抓取面部图像，以及未经同意创建面部识别数据库。
此外，法案还将对复杂和广泛使用的人工智能模型进行约束，要求它们生成的内容要有明确的标签，以解决虚假信息和选举干预的担忧。对于高风险人工智能系统，如自动驾驶汽车、医疗设备等，将实施严格的透明度义务。
欧盟计划在27个成员国中建立国家监管机构，以确保《人工智能法案》在整个欧盟范围内的统一执行。对于违规行为，企业可能会面临高达3500万欧元的罚款，或全球年营业额7%的罚款，具体取决于哪个数字更大。
全球影响方面，欧盟的《人工智能法案》被视为世界上第一个此类法案，可能会影响许多企业，尤其是那些开发人工智能系统的企业。科技巨头如Meta已经限制了其在欧洲的人工智能模型可用性，尽管这不一定是因为欧盟《人工智能法案》。
科技公司如Salesforce表示，欧盟的基于风险的监管框架有助于鼓励创新，同时也优先考虑安全开发和部署技术。金融科技公司Linedata的全球资产管理主管Jamil Jiva指出，欧盟通过巨额罚款确保监管产生影响的做法值得其他国家借鉴。
总之，欧盟《人工智能法案》的生效是全球范围内对人工智能监管的重要进展，旨在通过法律手段解决人工智能带来的风险，并确保其技术的负责任使用。
英语如下：
News Title: “EU AI Act Officially Comes into Force, Global Firms Face New Regulation”
Keywords: Act Comes into Force, EU Regulation, Artificial Intelligence
News Content:
Title: EU’s Artificial Intelligence Act Officially Becomes Effective, Marking the First Comprehensive Regulatory Framework for AI Globally
On Thursday, August 1, the EU’s Artificial Intelligence Act officially came into force, marking the first comprehensive regulatory framework for artificial intelligence globally and a significant step forward in Europe’s efforts to regulate the rapidly developing applications of AI. The implementation of the act will proceed in stages, aiming to provide businesses and systems with necessary transition periods to make the required adjustments and modifications.
The EU’s Artificial Intelligence Act employs a risk-based approach to regulate AI, which means that different applications will be subject to varying degrees of regulation based on the level of risk they pose to society. The act will start enforcing a series of rules beginning in February 2025, banning AI applications from exploiting personal vulnerabilities, prohibiting the non-targeted collection of facial images, and the creation of facial recognition databases without consent.
Moreover, the act will impose constraints on complex and widely used AI models, requiring them to provide clear labels for the content they generate, addressing concerns over false information and election interference. For high-risk AI systems such as autonomous vehicles and medical devices, strict transparency obligations will be implemented.
The EU plans to establish national regulatory bodies in its 27 member states to ensure the uniform enforcement of the Artificial Intelligence Act across the entire EU. Firms that violate the act may face fines of up to €35 million or 7% of their global annual turnover, whichever is greater.
On a global scale, the EU’s Artificial Intelligence Act is considered the first of its kind, potentially affecting numerous companies, particularly those involved in developing AI systems. Tech giants like Meta have already limited the availability of their AI models in Europe, although this is not necessarily due to the EU’s Artificial Intelligence Act.
Technology companies like Salesforce argue that the EU’s risk-based regulatory framework helps to foster innovation while also prioritizing the safe development and deployment of technology. The Global Asset Management Director at Linedata, Jamil Jiva, points out that the EU’s approach of ensuring regulatory impact through substantial fines is worth emulating by other countries.
In conclusion, the official coming into force of the EU’s Artificial Intelligence Act represents a significant advancement in global AI regulation, aiming to address the risks posed by AI through legal means and ensure its responsible use.
【来源】https://www.cls.cn/detail/1751761
Views: 28