Caifinance News on November 5th (Editor Xiao Xiang): Meta announced on Monday that it will allow United States government institutions and contractors involved in national security work to use its ai model Llama for military purposes. This undoubtedly changes the company's previous policy of prohibiting the use of its AI technology in such areas.

Meta stated that it will provide its ai model named Llama to federal agencies and is collaborating with defense contractors like Lockheed Martin, as well as technology companies focused on defense like Palantir and Anduril.

The Llama model is open source, which means that developers, companies, and governments worldwide can freely replicate and distribute the model technology. However, Meta's previous policy prohibited developers from using Llama for any projects related to military, war, or espionage tasks.

Currently, Meta is making an exception for the US defense agencies and contractors, aiming to expand its influence in the public sector and increase Llama's foothold in the emerging AI arms race.

Meta's Global Affairs Chief, Nick Clegg, stated in an announcement on the company's official website on Monday that the company now supports the "responsible and ethical use" of this technology. Clegg believes that the responsible and ethical use of open source AI models like Llama will not only support America's prosperity and security but also help establish American open source standards in the global AI leadership competition.

"Meta hopes to contribute to the security, safety, and economic prosperity of the United States and its closest allies," said Clegg. "Widespread adoption of American open source AI models is in line with both economic and security interests."

A spokesperson for Meta stated that the company will share its technology with other members of the Five Eyes alliance. In addition to the USA, the Five Eyes alliance countries include Canada, United Kingdom, Australia, and New Zealand.

Meta, owning popular social applications such as Facebook, Instagram, and WhatsApp, has been actively promoting its large-scale artificial intelligence models to as many third-party developers as possible amidst the competition from rivals like OpenAI, Microsoft, Google, and Anthropic. In a different strategy from the aforementioned main competitors, Meta, which started behind in the field of artificial intelligence compared to these companies, decided to open-source its code to catch up.

According to a recent study by the non-profit organization AI Now Institute, the use of artificial intelligence in military intelligence, surveillance, and reconnaissance poses dangers as it relies on personal data, which could be stolen and weaponized by adversaries. The co-authors of the study wrote that AI also has weaknesses such as bias and creating illusions, with no remedies currently available, and they suggest creating artificial intelligence independent of 'commercial' models.

However, Meta asserts that open AI can accelerate defense research while promoting the 'economic and national security interests' of the USA. Clegg wrote that in defense and security fields, Llama can assist in tasks like data analysis, document synthesis, or 'tracking terrorist financing or strengthening our network defenses.'

Currently, progress in the adoption of AI technology by the US military remains slow, with skepticism towards its return on investment. So far, the US Army is the only branch of the US armed forces deploying generative AI.

It is worth mentioning that during last week's company earnings call, Meta's CEO Zuckerberg also mentioned that he is working to get 'government use of Llama across the entire US government.' The latest updates may help achieve this goal, paving the way for Zuckerberg to further develop relationships with elected officials.