在2024年国际消费电子产品展（CES 2024）上，知名安全软件公司McAfee发布了一项名为“Project Mockingbird”的新人工智能技术。这项技术旨在通过深度学习模型分析视频内容，并辨别出其中是否可能被深度伪造技术（Deepfake）所操纵的音频。据初步测试结果显示，McAfee的这一检测系统在识别和区分人工生成的，冒充真实人的虚假音频方面的准确率超过90%。
随着深度伪造技术的不断进步，伪造的音频和视频内容变得越来越难以辨认。这种技术可以用来伪造音频，使其听起来就像真实的某人所说的话，这可能会被用于网络钓鱼、误导信息传播和其他恶意目的。因此，开发能够有效检测此类伪造内容的技术变得至关重要。
“Project Mockingbird”利用机器学习算法来训练模型，识别异常模式和声音特征，从而判断音频是否经过了深度伪造处理。这种技术的推出，不仅展示了McAfee在网络安全领域的深厚实力，也体现了人工智能技术在应对现代网络安全挑战中的应用潜力。
McAfee表示，他们将继续改进“Project Mockingbird”，并计划在未来的产品中集成这一功能，以帮助个人和企业更有效地防御对抗深度伪造攻击。
英文翻译：
Title: McAfee Unveils “Project Mockingbird” AI to Detect Deepfake Audio
Keywords: McAfee, AI technology, Deepfake audio detection
News content:
At CES 2024, the renowned cybersecurity company McAfee has launched a new artificial intelligence technology dubbed “Project Mockingbird.” This technology aims to analyze video content to detect whether audio may have been manipulated by deepfake techniques. Preliminary test results indicate that McAfee’s detection system achieves an accuracy of over 90% in identifying artificially generated fake audio that impersonates real people.
As deepfake technology advances, fabricated audio and video content has become increasingly indistinguishable. This technology can be used to forge audio, making it sound like a real person, which could be employed for phishing, spreading misinformation, and other malicious purposes. Therefore, developing technology that can effectively detect such content is critically important.
“Project Mockingbird” employs machine learning algorithms to train models that recognize anomalous patterns and sound features, thus determining if the audio has been subject to deepfake manipulation. The introduction of this technology not only showcases McAfee’s strong expertise in cybersecurity but also highlights the potential application of artificial intelligence in addressing modern cybersecurity challenges.
McAfee plans to continue refining “Project Mockingbird” and integrate it into their future products to help individuals and enterprises more effectively defend against deepfake attacks.
【来源】https://www.ithome.com/0/744/054.htm
Views: 1