US signals cooperation with China on AI safety: what could happen, and what could it mean?

Two significant events occurred between the executive order issued in August 2023 and the statement by the White House’s top science adviser. Firstly, in November, China joined 27 other countries and the European Union to sign the Bletchley Declaration on evolving standards for AI at the world’s first AI Safety Summit in the UK. Secondly, later that month, President Biden met with his Chinese counterpart Xi Jinping at the APEC summit for discussions including on US export controls for sensitive equipment.

Why is the statement by the White House important?

• The shift in Washington’s approach comes amid global concerns that rapid advancements in AI could increase vulnerabilities to cyber attacks and misinformation, especially as a record number of countries are set to vote in elections this year.
• A recent Forbes report quoted analysts, including Kaifu Lee, stating that the US and China have reached parity in AI development, but China’s implementation of the technology in products and services is likely to surpass the US this year.
• Chinese AI major Baidu, in collaboration with Alibaba and Tencent, has recently introduced ERNIE, a 23-billion-parameter AI model trained on nearly 150 million Chinese image-text pairs.
• Another Chinese AI model, Taiyi, is a bilingual (Chinese-English) large language model trained on about 20 million filtered Chinese image-text pairs and one billion parameters.
• China aims to become the global leader in AI by 2030, with TikTok seen as a global leader in behavioral algorithms. China is also advancing in facial recognition, outpacing the US and Europe, due to the heavy deployment of a billion surveillance cameras in Chinese urban centers that are integrated with backend AI tools.

How will a collaborative approach towards China work for the US?

• While the specifics of the new collaboration between the US and China on AI are unclear, Arati Prabhakar, director of the White House Office of Science and Technology Policy, stated that there are areas of agreement between the two countries, including global technical and safety standards for AI software.
• However, she emphasized that the US does not intend to slow down AI development but aims to maintain oversight of the technology.
• This comes amid criticism that the Biden administration’s early regulation of AI could have negatively impacted the competitive advantage of American tech companies.
• Prabhakar argued that even American AI companies acknowledge the need for clear methods to understand and assess AI, but the best options for evaluating the safety of new AI systems are currently inadequate due to the complex and opaque nature of the technology.
• Despite the US imposing export restrictions on advanced chips like Nvidia’s latest GPUs to limit China’s ability to train foundational AI models, there is a recognition in Washington that the US and Chinese AI research ecosystems are deeply interconnected.
• Some of the best AI scientists from China come to the US, and Washington can benefit from these ties to China, including the flow of talent.

What could be the impact of this seeming reconciliation in other geographies?

• A US-China alignment on AI technology, particularly its regulation, could influence other regions, including India. India has been positioning itself as a nation that effectively uses technology to develop and deliver large-scale governance solutions, such as the Aadhaar biometric identity program and the Unified Payments Interface (UPI).
• These solutions form the foundation of what New Delhi terms Digital Public Infrastructure (DPI), where the underlying technology is approved by the government and then offered to private entities to develop various use cases. India aims to apply this DPI approach to AI.
• With sovereign AI and an AI compute infrastructure, New Delhi aims not just to compete with the generative AI model, but also to focus on real-life use cases in healthcare, agriculture, governance, language translation, etc., to maximize economic development.
• Currently, the European approach to technology rules is seen as inherently regulatory for citizens’ rights, while the US approach is more focused on innovation. According to policymakers, India’s approach has been a hybrid of the European and American approaches.
• The new US-China consensus on tech development and regulation could provide another useful model for India.

Conclusion- The recent alignment between the US and China on AI technology and its regulation marks a significant shift in global tech dynamics. This collaboration could have far-reaching implications, particularly for countries like India that are leveraging technology for large-scale governance solutions. However, it’s crucial for India to maintain a balanced approach, considering both citizens’ rights and innovation, as it navigates its path in the rapidly evolving AI landscape. The interplay of geopolitics and technology will continue to shape the future of AI, with potential impacts on cybersecurity, misinformation, and economic development.