China uses Meta’s Llama language model to develop military AI

Researchers at Chinese universities have developed a language using Meta’s Llama open source language model. artificial intelligence tool for military activitiesas explained by Reuters. The platform was called “ChatBIT”.

The technology, based on Llama 13B, is capable of collecting and processing intelligence data after the inclusion of certain parameters. In this way, it can provide information that will help decision-making, as seen in the articles obtained by the news agency.

Optimized for performing dialogue tasks and answering questions in the military ChatBIT reportedly outperforms powerful generative AI models Like OpenAI’s ChatGPT-4. Yet only 100,000 soldiers were trained in dialogue; this is a much smaller data set than those normally adopted in other intelligent systems.

In the future, it is possible that the vehicle will be used in tasks such as strategic planning, simulations and command decisions, as detailed in the texts. However, the documents do not indicate whether the smart bot is already used by China.

Meta bans use in military projects

Amid the terms of use of Meta’s open source model is a prohibition on using the technology in the development of systems for “military, combat, nuclear, and espionage industries or applications.” However, the company has difficulty controlling this.

The company questioned by the publication said that the use of artificial intelligence in the way described in scientific articles is unauthorized and against their policies. He also emphasized that China has invested more than US$1 trillion (R$5.87 trillion) in artificial intelligence to surpass the US, and suggested that the country has much more advanced models than ChatBIT.

The tool was created by research institutions affiliated with the Chinese People’s Liberation Army, such as the Academy of Military Sciences. Studies on the bot were revealed in June.