In the aftermath of the recent meeting between President Joe Biden and Chinese General Secretary Xi Jinping, a bilateral channel has been established between the two nations. The Brookings Institution reported that this channel provides both governments with an opportunity to mitigate potential risks and set up governance for AI and future technologies.
The Brookings Institution's article by Graham Webster and Ryan Hass stated that during the November 2023 meeting in California, both governments unveiled a new bilateral consultation channel on artificial intelligence (AI). At present, the term "AI" does not have a specific definition. However, according to Webster and Hass, for any progress to be made between the United States and China, it is necessary to address certain specific topics.
Webster and Hass said that both countries are well aware of the implications of AI models. This new channel, which was introduced last year, facilitates detailed discussions about AI. Given that companies from both the United States and China are at the forefront of AI innovation, there is an urgent need for agreements and governance as competition intensifies and trust diminishes. They urged government officials to concentrate on three main areas: military applications of AI, fostering positive cooperation between the United States and China, and focusing on achievable goals.
Webster stated that both nations have incorporated AI into military systems such as autopilot and computer vision. The objective is not to completely avoid military AI but rather to establish boundaries and shared expectations for acceptable automation. This would involve setting testing standards before deploying AI-enabled systems along with early agreements like affirming that only humans should make nuclear launch decisions; automation in such matters would be considered unacceptable.
The Brookings Institution's article also underscored the importance of trust amidst common challenges between the United States and China. Both governments share concerns about how to label outputs from AI models. Webster suggested that teams from both countries should collaborate to enhance compatibility in these systems, possibly involving industry stakeholders and global standards organizations for a more comprehensive strategy.
The authors pointed out that the main challenge of the new AI channel between the United States and China is the risk of becoming overly entangled with unresolvable issues. Examples include China's dissatisfaction with the United States' export controls on advanced GPUs and the United States' concerns about China's domestic surveillance. Instead, they suggested both sides should focus on practical governance and seek early successes to demonstrate the value of the channel.
According to its website, the Brookings Institution is a nonprofit organization that strives to improve all levels of governance through nonpartisan research. Hass holds a senior position at the institution and also serves as a director.
