Chinese researchers have developed a military-optimized chatbot named ChatBIT using Meta's open-source Llama 13B artificial intelligence model, according to recent reports.
The project involves six researchers from three institutions, including two affiliated with the People's Liberation Army (PLA). They utilized a version of Llama to create an AI tool aimed at processing military information and enhancing decision-making capabilities.
ChatBIT reportedly surpasses other AI models, achieving performance levels close to OpenAI's ChatGPT-4. The researchers have not disclosed specifics on how they measured performance or whether the AI is currently operational.
Meta has restrictions against military use of its models, emphasizing that any such application is unauthorized. Despite this, the open-source nature of Llama limits Meta's ability to enforce these restrictions effectively.
This development occurs amid heightened concerns in the U.S. regarding AI technologies and national security, prompting the Biden administration to finalize regulations aimed at limiting U.S. investments in Chinese AI sectors.
In response to these advancements, Pentagon officials acknowledged the dual-use nature of open-source models and the need for ongoing monitoring of competitor capabilities.