“Deepfakes” and artificial intelligence technology are being used by foreign forces to attack China, according to an ominous warning issued by China’s top intelligence agency.
In a rather alarming move, the Ministry of State Security sounded the alarm bell on Friday, citing the misuse of Artificial Intelligence technology, where outsiders have developed deceitful videos to manipulate public opinion and spark widespread panic in China. It is rather alarming to observe the rapid misuse of the latest advancements in Artificial Intelligence.
“The technology uses complex AI to create highly believable but purely synthetic images, sound, and video,” the MSS said, while also pointing out that although such technology has certain uses in the fields of entertainment and education, “it lends itself to highly damaging activities when pursued by those seeking to attack the fabric of our society and our national security.”
“Certain anti-China foreign forces have utilized this technology to create fake videos, and attempted to spread such videos within China in an aim to deceive the masses and cause panic,” said the security bulletin from the foreign ministry.
In addition to the development of fake news through technology, the intelligence organization identified loopholes within how sensitive information is handled by AI technology.
Beyond the Screen: China’s MSS Warns of Deepfakes, Data Leaks, and Algorithmic Bias in AI
According to the ministry, there was a disturbing instance where its employees had used open-source platforms of AI technology without proper security measures to prevent anyone from accessing and downloading sensitive information through foreign IP addresses without even needing a password.
But the warning also highlighted another issue, albeit a more subtle and potentially pervasive one: algorithmic bias.
As many advanced language models have been trained on data that sometimes reflects existing biases and prejudices of society, they have the potential to perpetuate and further entrench discrimination and inequities and underrepresent the viewpoints of marginalized groups, according to the ministry.
The MSS issued useful tips to help Chinese citizens and institutions safeguard themselves against this rapidly changing threat environment.
The key takeaways are to use AI applications extremely carefully before allowing them to gain access to your personal information, avoid online AI utilities while processing sensitive information, and always keep payment password information out of AI systems.
It is also important for users to turn off data sharing functionality and cloud backup solutions if available, according to the ministry. It is critical to continuously watch their “digital footprints” online trail created by their online activities and delete chat conversations with their AI assistants occasionally, aside from updating their antivirus and maintaining a healthy level of skepticism against demands for their personal and financial identities.
“Security is the prerequisite for development, and development is able to guarantee security,” explained the MSS, defining the importance of cybersecurity and China’s further development through these lines: “Only by understanding technology and using it safely, can we make sure that large-scale AI models could become a positive force driving forward social progress.”
China’s Multi-Layered Strategy to Combat AI Exploitation
It is not the first instance that has come to light wherein the intelligence agencies in China have raised red flags regarding the risks associated with vulnerabilities in the use of AI.
In March itself, the ministry had appealed to both companies and government organizations to “prevent the leak of data” during the training procedures for artificial intelligence models.
The concern that had been raised was about “criminals taking advantage of system weaknesses to attack smart systems in factories, banks, and power plants that could lead to both economic losses and adverse physical incidents.”
The government has followed these threats with legislative measures. Last October, China has adopted amendments to the Cyber Security Law, of which a completely new chapter exclusively focuses on AI security and development.
The measures outlined in the revised law include higher ethical standards in AI usage, better monitoring of security risks for AI, and stronger regulation of AI usage in every sector.
The MSS ended its latest warning by urging citizens to report any fishy AI behavior, especially when models are suspected of data-stealing and when they are transmitting private data abroad.
The crowd-sourced manner in which the threat is being detected is an indication of the problem that the Chinese government is currently dealing with in terms of AI evolving faster than its security frameworks.
With artificial intelligence intertwining in an increasingly deeper way in the fabric of life all over the world, China’s alertness is a wake-up call regarding the fact that these powerful tools involve major security considerations. It may be in the form of major countries using deepfakes for malicious purposes, or in the form of simple security lapses in which data leaks out because of poorly implemented AI tools.
