After allegedly stealing fighter jet technology to speed up its aircraft manufacturing program, China may have done it again. This time, it has reportedly exploited the publicly available Meta’s AI tool, the Llama model, to develop its own military applications.
A perturbed Meta has now brought a policy change to allow the US government agencies and contractors working on national security to use its AI models for military purposes.
Top Chinese research institutions with ties to the People’s Liberation Army (PLA) have used the source code of Meta’s AI tool as the base for its military intelligence tool, ChatBIT. Six Chinese researchers from three institutions under the aegis of China’s leading research institution, the Academy of Military Science, wrote a paper on it.
The researchers used Meta’s earlier Llama 213B large language model (LLM), which Meta had made public. The Chinese researchers incorporated their parameters to construct a military AI tool to gather and process a large amount of information. The AI tool would then facilitate faster operational decision-making.
The Llama models are “open-source,” which means the technology can be freely copied and distributed by other developers, companies, and governments. It is the first time there has been irrefutable evidence that the PLA military experts have been systematically researching and trying to leverage the power of open-source LLMs, especially those of Meta, for military purposes.
The Chinese research comes as there has been a raging debate in the US national security and technology circles on whether AI models should be open source. US President Joe Biden signed an executive order in October 2023 to manage AI developments.
The paper said that ChatBIT was fine-tuned and “optimized for dialogue and question-answering tasks in the military field.”
In terms of accurate and reliable information, ChatBIT outperformed 90 percent of other Open AI, including ChatGPT-4. The Chinese application was used for about 100,000 military dialogues, which is considered far less than for typical big language models.
This has forced Meta to change its policy. Until now, Meta’s approved use policy forbade applications connected to war, nuclear industries, or espionage. It also prohibited Meta AI from military usage. Molly Montgomery, Meta’s head of public policy, told Reuters, “Any use of our models by the People’s Liberation Army is unauthorized and opposite to our acceptable use policy.”
However, Meta is now ready to make its AI model, Llama, available to federal agencies and defense firms like Lockheed Martin. This move is an exception to Meta’s “acceptable use policy.”
In a blog post, Nick Clegg, Meta’s president of global affairs, said the company now backed “responsible and ethical uses” of the technology that supported the United States and “democratic values” in a global race for AI supremacy.
“Meta wants to play its part to support the safety, security, and economic prosperity of America — and of its closest allies too,” Clegg wrote in the post. “We believe it is in both America and the wider democratic world’s interest for American open-source models to excel and succeed over models from China and elsewhere,” he added.
The AI tools can process vast amounts of data and generate usable insights. Large language models can help to streamline complicated logistics and planning, track terrorist financing, or strengthen cyber defenses.
For decades, open-source systems have helped build the world’s most technologically advanced military. They have also helped accelerate defense research and high-end computing, identify security vulnerabilities, and improve communication between disparate systems.
Now Meta is partnering with companies including Accenture Federal Services, Amazon Web Services, Anduril, Booz Allen, Databricks, Deloitte, IBM, Leidos, Lockheed Martin, Microsoft, Oracle, Palantir, Scale AI and Snowflake to bring Llama to government agencies.
Oracle is building on Llama to synthesize aircraft maintenance documents so technicians can more quickly and accurately diagnose problems, speeding up repair time and getting critical aircraft back in service.
Scale AI is fine-tuning Llama to support specific national security team missions, such as planning operations and identifying adversaries’ vulnerabilities. Lockheed Martin has incorporated Llama into its AI Factory, accelerating code generation, data analysis and enhancing business processes.
Clegg contends that national security is inextricably linked with economic output, innovation, and job growth. The widespread adoption of American open-source AI models serves both economic and security interests. Noting this, China is racing to develop its open-source models and investing to leap ahead of the US.
Meta has also argued for collaborating with states that have committed to the Political Declaration on Responsible Military Use of Artificial Intelligence and Autonomy.
AI In Military
AI has revolutionized military technology. It has versatile uses in autonomous weapon and vehicle systems, intelligence command and control systems, predictive maintenance, Logistics and maintenance, cybersecurity, intelligence and surveillance, decision-making applications, and simulations and training.
AI will introduce a new element to warfare: supplementing and augmenting human cognition. Militaries aided by AI will be able to operate faster and conduct more complex operations.
While humans can only pay attention to a smaller number of things, an AI-enabled system can absorb and process a large amount of information at once. Machines can process data faster than humans and guide military units to appropriate responses.
This would not make humans redundant. Humans will still be needed in real-world combat environments to deal with the unknowable.
