在2024年CES展会期间，知名安全软件公司麦咖啡（McAfee）推出了一项名为“Project Mockingbird”的新技术。这项创新的人工智能解决方案旨在分析视频内容，并有能力辨识出是否含有经深度伪造技术处理的音频。据早期测试结果显示，麦咖啡的这一检测系统在识别由人工智能生成的、冒充真实人的声音方面的准确率超过90%。
深度伪造技术，或称Deepfake，近年来引发了广泛的社会关注和安全隐患。这种技术可以用来制作极其逼真的音频、视频内容，用于欺骗、诽谤、侵犯隐私或其他非法目的。随着技术的不断进步，Deepfake音频的检测变得越来越重要，尤其是在保护个人隐私和维护网络安全方面。
“Project Mockingbird”通过复杂的机器学习模型，能够识别和分析音频信号中的异常模式，从而判断音频是否经过了深度伪造处理。这项技术可能会对社交媒体平台、政府机构、金融机构以及任何需要保护自身免受伪造信息侵害的组织产生重大影响。
麦咖啡公司表示，他们将继续改进“Project Mockingbird”，以应对不断演变的伪造技术。未来，该技术也可能扩展到视频领域，提供全方位的深度伪造检测服务。
英文标题：McAfee Unveils “Project Mockingbird” AI to Detect Deepfake Audio
英文关键词：McAfee, AI detection, deepfake audio
英文新闻内容：
At the 2024 CES event, the renowned cybersecurity company McAfee has launched “Project Mockingbird,” a groundbreaking AI-driven solution designed to analyze video content and has the capability to detect whether audio contains manipulations through deepfake technology. Early testing results indicate that this detection system by McAfee achieves over 90% accuracy in identifying artificial, human-like voices generated by AI.
Deepfake technology, or deepfakes, has sparked widespread societal concern and security risks in recent years. This technology can be used to create极其逼真的audio and video content for purposes of deception, defamation, privacy invasion, or other illegal activities. With the continuous advancement of technology, the detection of deepfake audio has become increasingly critical, particularly in safeguarding personal privacy and maintaining network security.
“Project Mockingbird” employs complex machine learning models to identify and analyze anomalies in audio signals, determining whether the audio has been subject to deepfake manipulation. This technology could have a significant impact on social media platforms, government agencies, financial institutions, and any organizations needing to protect themselves against misinformation and fake content.
McAfee states that they will continue to refine “Project Mockingbird” to adapt to the ever-evolving deepfake landscape. In the future, the technology may also expand into the video domain, offering comprehensive deepfake detection services.
【来源】https://www.ithome.com/0/744/054.htm
Views: 1