AI technology has shown great potential to spark technological change and improve productivity and efficiency, but at the same time, the security risks it poses are becoming more and more prominent. The use of deepfakes to create illusions, manipulate public opinion, and even carry out fraudulent activities has intruded into normal public life, and new methods of "AI scams" have emerged one after another, making it difficult to prevent them, and AI technology has become an "accomplice" of new telecom fraud.
At the "3.15" party a few days ago, the information technology laboratory announced a series of "new scams" using AI technology: a woman received a call that her daughter had been kidnapped, and the cry of her daughter on the other end of the phone made the lady panic. After calling the police, the police investigation proved that the call was a fraud by fraudsters borrowing a virtual number and then using AI onomatopoeia technology. Another case seems to be more "advanced", the company employee received a video call from the boss and was asked to transfer more than 1.8 million yuan, but in fact, the "boss" in the video is just a "Li Ghost" made with AI synthesis technology.
Deepfakes are not a new technology, but today, the explosive development of generative AI technology has greatly lowered its barrier to entry and enhanced the realism of counterfeiting.
Behind these cases, criminals used AI technology to imitate people's facial images and voices. What the victim sees is not necessarily believing, and what he hears may also be false.
Police pointed out that whether it is an AI simulated voice or an AI video call, its duration is generally short, usually only a few seconds. This is mainly due to the fact that these video calls are pre-made fake videos, and the longer they are played, the more likely they are to be seen by the victim. At the current state of technology, it is extremely challenging to achieve highly realistic AI face swapping in peer-to-peer live video calls. This not only requires a breakthrough in the security protection system of video calling software, but also requires the input of a large amount of data, the support of professional algorithms, and continuous optimization and iteration.
01
The AB side of AI technology
Artificial intelligence is a strategic technology that leads a new round of scientific and technological revolution and industrial transformation. On the one hand, artificial intelligence empowers national security-related fields and is a powerful starting point for building a solid national security barrier. On the other hand, due to its vulnerability, instability, unexplainability and other characteristics, artificial intelligence is very likely to cause security risks at the national, social, enterprise and individual levels in the process of deep integration and application with the economy and society.
AI technology is like a coin placed in the hand, no matter how "eye-catching" the upward side is, you can't ignore the existence of the reverse side. With the opening and application of large models represented by ChatGPT, the threshold for the use of AI-generated technology has been continuously lowered, which has aroused widespread concern about AI security and compliance issues from all walks of life. According to a report by cybersecurity firm Darktrace, ChatGPT-like generative AI technology has led to a 135% spike in phishing email attacks.
In addition, the use of deep learning technology to "deepfake" means can realize the tampering and falsification of multimedia content such as images, sounds, and videos, and generate highly realistic and unrecognizable false information, which seriously interferes with the public's cognition and judgment ability. It is very common to search for AI face swapping and AI voice changing on the Internet, and it is very common to enter a picture to easily "swap faces", and you can also freely train and formulate voices to synthesize video output. Not only that, Qi Xiangdong, chairman of Qianxin Technology Group, previously pointed out that generative AI technology can also quickly generate phishing emails, write malware and code, resulting in frequent AI fraud incidents.
02
"Defeat magic with magic"
AI治理AI
With the increasing emphasis on AI security governance around the world, technologies in the field of security are also constantly innovating.
In this regard, Qi Xiangdong believes that "the security problems brought by AI can only be solved by AI. "On the one hand, AI makes security issues prominent, and at the same time, it also empowers the cybersecurity industry and plays a great role in promoting the upgrading of offensive and defensive technologies, which will promote the competition between attack and defense in the use of AI.
"AI will not replace cybersecurity experts, but it will make cybersecurity experts more capable of dealing with danger, so AI should set off a technological revolution in the field of cybersecurity and make the cybersecurity industry better. Qi Xiangdong said that leading companies in various industries should be encouraged to cooperate with cybersecurity companies to integrate AI security technology into digital scenarios for effective security protection.
Zhou Hongyi, founder of 360 Group, suggested that the industry should formulate a standard system to ensure the security of general large models, promote general large models to carry out security evaluation, access security services, and reduce the security risks of general large models. Turing Award winner Joshua Bengio pointed out that science cannot yet ensure that these future AI systems are safe, and that scientific solutions and global political solutions should be found now.
From the supply side, we will develop cutting-edge technologies of "AI + security" around offensive and defensive application scenarios. Encourage leading enterprises in various industries to cooperate with professional security vendors to explore the application of large-scale model security innovation products in threat detection, vulnerability mining, command and judgment, etc., and promote "AI + security" into a virtuous circle.
From the demand side, it is suggested that the government should strengthen policy traction, support enterprises that develop and innovate "AI + security" products, promote "AI + security" related industries to achieve more scientific and technological innovation achievements, and expand the field of "AI + security".
END
Author: Wang Hejia Editor/Layout: Gai Beibei Reviewer: Wang Tao Mei Yaxin Producer: Liu Qicheng
【通信世界新媒体矩阵】央视频 | 微软MSN | 视频号 | 微博 | 今日头条 | 百家号 | 网易号 | 搜狐 | 腾讯新闻新浪看点 | 雪球号 | 抖音 | 快手 | 爱奇艺 | 知乎 | 哔哩哔哩 | 咪咕视频 | CSDN | 36氪
【New Media Team】Producer|Liu Qicheng Proofreader|Wang Tao, Mei Yaxin Editor|Wang Yurong, Gai Beibei Video Editor|Huang Yangyang, Lu Ruixu Operation|Lin Song
Dots to share
Point Collection