The U.S. National Security Commission on Artificial Intelligence released its final report recently, listing China as a strategic competitor to the United States in this field. The report describes China as a U.S. peer in many areas and an AI leader in some areas. This new technology allows machines to exhibit characteristics associated with human learning and problem-solving, and can be applied to areas such as facial and speech recognition, natural language processing, and automated reasoning.
While China has made technological strides in the AI field, the authors of the report view these developments as a threat. As recorded in the report, potential threatening applications can be made in a number of areas.
First, AI boosts the threat imposed by potential cyberattacks coming from China. Cyberattacks can be made more rapidly, with better precision, and in greater secrecy with the use of AI. Already, cyberattacks have been used to steal trade and government secrets. Intellectual property protection was a central issue in the China-U.S. trade war and may become more vulnerable as China accelerates its AI capabilities. Cyberattacks have also been used to disseminate disinformation, which was prevalent during the 2016 U.S. election, and spread self-replicating AI-generated malware. Use of AI-fused data for blackmail, deepfakes, or swarms are possible in the future.
Second, China plans to use AI to offset U.S. military superiority by implementing a type of “intelligentized war” that relies more on creation of alternative logistics, procurement, and training, as well as warfare algorithms. Battle networks will connect systems, and armed drones with autonomous functions will be employed. Soldiers will be trained in live and virtual environments that integrate AI. AI will speed up the process with which valuable targets can be identified and hit due to enhancements in collection and transmission of intelligence.
Third, China’s use of AI in national intelligence will help government officials pinpoint trends and threats as well as use deception and expose sources and methods. AI renders social media information, satellite imagery, communications signals, and other sources of data more understandable and potentially actionable. Intelligence sources may be coupled with domestic and international surveillance. The authors assert that “China’s domestic use of AI is a chilling precedent for anyone around the world who cherishes individual liberty” due to its use in domestic surveillance and repression.
To combat these possibilities, the report emphasizes that the United States should ensure that it also builds its own AI capabilities and not fall behind. The authors recommend that the U.S. invest $40 billion in expanding and democratizing federal AI research and development. They also recommend that the U.S. create a Joint Interagency Task Force and Operations Center and provide additional funds to the Defense Advanced Research Projects Agency (DARPA) to counter social media disinformation. As a response to hacking and other attacks, the U.S. should develop AI-enabled defenses against cyberattacks and set up red teams for adversarial testing. The report also recommends that, in order to maintain military defense capabilities against AI-based attacks, the Department of Defense should invest in next-generation technologies and set up a joint warfighting network architecture this year.
Certainly, the world of AI will lead to the danger of automating and accelerating decisions that can harm other nations. One of the biggest issues is that AI-based applications may automatically authorize use of nuclear weapons. This emphasizes the need to ensure that AI cannot make security-critical decisions without some human intervention. As long as humans are somehow involved in the final decision, possessing the capability to engage in cyberattacks, intelligentized war, or national intelligence gathering does not in itself place other nations in direct jeopardy. Equally importantly, ensuring that there are open channels to negotiate disputes and keep the peace is the most critical aspect of reducing conflict.
It is important for world powers to maintain AI capabilities in the event of a conflict, but these offensive AI-based technologies should not be the first response for the U.S. or any other nation. This can only serve to escalate conflict and increase the likelihood that two or more nations will engage in a broader war. Reliance on AI-based conflict should be viewed as a last resort, and diplomatic and economic relations should be used as the primary method of maintaining peace.