China’s intelligence chief just gave a revealing glimpse into what Beijing fears most about AI

China's spy main conscionable offered a model into Beijing's priorities erstwhile it comes to AI's biggest risks.

Chen Yixin, China's curate of authorities security, published an article Sunday greeting successful China Cyberspace, a diary tally by the Cyberspace Administration of China, the country's net regulator.

In it, Chen identified six awesome risks AI could airs to Beijing. The first was "regime security."

He wrote that "various dispute forces and malicious individuals maltreatment generative AI technologies — specified arsenic deepfake audio and video, AI-generated images and text, and intelligent online armies — to nutrient governmental rumors astatine debased costs and successful ample quantities."

Those devices tin "spread harmful information, incite antagonistic emotions, and costs a warfare of nationalist sentiment and cognitive warfare," posing nonstop threats to China's political, institutional, and ideological security, he said.

Chen besides identified 5 different awesome AI risks, including:

• Dangers to China's captious infrastructure, because cyberattacks tin beryllium carried retired pinch utmost efficiency, considering the powerfulness of models for illustration Mythos.
• The consequence of large-scale information leaks, wherever Chinese AI users and overseas intelligence agencies alike could get captious nationalist data, waste and acquisition secrets, aliases citizens' individual information.
• An imbalance successful technological development caused by monopolies and closed technological systems, arsenic countries tin utilization AI's powerfulness and part proviso chains successful the sanction of nationalist security.
• Challenges to societal governance that lead to disruptions successful nationalist order.
• Fundamental shifts successful warfare, arsenic AI-powered precision targeting gives adversaries a awesome advantage connected the battlefield.

More regularisation for the 'wellbeing of each humanity'

Chen advocated for beardown "developmental guidelines" for AI to beryllium put successful spot — essentially, much authorities power of AI development.

What that involves, per Chen, will beryllium for Beijing to "accelerate the building of a information consequence prevention and power system." China should beforehand the "healthy and orderly development" of AI done "high-efficiency governance," Chen said.

Chen besides said AI should beryllium developed "with the wellbeing of each humanity arsenic its guiding principle."

He added, too, that world powers must "resolutely defy hegemonism, technological barriers, and exclusive blocs," and activity to group up a "fair and unfastened world AI rules system."

Chen's comments travel arsenic Chinese leader Xi Jinping advocated for AI progress astatine the BRICS acme successful New Delhi connected Sunday.

"We should return a people-centered approach, create AI for the affirmative and for good, and activity faster to germinate a consensus-based world AI governance framework, truthful that caller technologies tin ray the measurement toward shared prosperity," Xi said.

The comments from Chen and Xi came arsenic the debate complete a imaginable AI apocalypse heats up successful the US. Over the weekend, Dario Amodei, Sam Altman, and Elon Musk each backed calls to slow the pace of frontier AI development.

The CEOs voiced these concerns aft respective of their employees issued nationalist warnings that the products they are making, if developed unchecked, could swipe retired astir of humanity.

In the astir precocious floor plan instance, Anthropic interrogator Jacob Coxon discontinue Anthropic, saying some the laboratory and ChatGPT-maker OpenAI are "gambling pinch our lives."