GET DAILY REPORT FREE
The U.S. and China are taking noticeably different approaches to artificial intelligence safety as AI systems become more powerful and capable of acting autonomously, The New York Times reports.
In the U.S., recent incidents involving AI systems hacking computer networks have increased anxiety and debate about how to manage the risks of advanced AI. American technology companies, employees and policymakers are increasingly discussing stronger safeguards, including proposals to slow development or create ways to shut down powerful systems.
China, however, has generally viewed AI as a manageable technology whose benefits outweigh its risks. Chinese companies also strongly support open-weight AI, which allows others to access and modify a model’s underlying parameters.
Supporters argue that openness can improve security by giving more people the opportunity to identify vulnerabilities and develop solutions.
At the same time, China’s approach to AI safety is changing as its companies approach the technological frontier. Chinese universities, government laboratories and technology companies have increased research into the safety of advanced AI models and AI agents. This represents a broader shift from focusing mainly on controlling AI-generated content to examining what AI systems can actually do, particularly their ability to act independently and create cybersecurity risks.
The Chinese government has also called for AI to remain under human control and has encouraged international cooperation on AI safety. Recent cybersecurity incidents involving U.S. AI systems have strengthened arguments in China that open models can provide useful defenses against unsafe technology.
The New York Times has the full story. A subscription may be required.
