Meta is trying to broaden its reach in the public sector and strengthen Llama’s position in the emerging AI arms race
Meta Platforms Inc. has authorised US government agencies and defence contractors to use its AI models, which opens the door for Meta’s technology to play a vital role in US military and national security.
The Facebook parent company is giving access to its language model Llama to more than a dozen US agencies and contractors, including Lockheed Martin Corp., Booz Allen Hamilton Holding Corp., and Palantir Technologies Inc., along with some tech industry leaders like Microsoft, IBM, and Amazon Web Services to use its AI technology.
Oracle is developing technologies to synthesize aviation maintenance records, intending to reduce the time to repair military aircraft. Scale AI is customizing its technology for the national security team for threat assessment and operational planning. Lockheed Martin used Llama for its AI Factory to improve its defense-related applications.
Although the Meta AI model is open source, which means the developers are free to use it, Meta’s acceptable use policy restricts people from using it for any initiatives that involve the military, armed conflicts, nuclear industries or applications, and espionage, among other things.
But Meta is trying to broaden its reach in the public sector and strengthen Llama’s position in the emerging AI arms race. It is making an exemption for the US defense agencies and their contractors.
Nick Clegg, the president of global affairs at Meta, states that the company would do its part to support the safety, security, and economic prosperity of the United States and its close allies. Meta is an American company, and it owes its success to the entrepreneurial spirit and democratic values of the country.
Meta gives access to similar government agencies and contractors in the UK, Canada, Australia, and New Zealand. These countries have an intelligence-sharing partnership called Five Eyes.
But Israel, which is the closest ally in the Middle East, is not included in the present list of partners.
The decision was taken just before the US election, and Meta is worried that developers in countries like China may use Llama for military purposes.
Clegg added that the present AI boom is based on language models and can be used to perform various tasks like powering chatbots. In the world of defense and security, Llama can help with many tasks like data analysis, document synthesis, tracking and funding for terrorism, or strengthening cyber defenses.
Although Meta will not control how US agencies or its partners use the Llama technology, it is the responsibility of the country that uses AI for natural security to use AI ethically and follow relevant international law.
Meta wanted to work with the US defense mainly because of its goal of establishing Llama as the leading AI product in the world.
Chief Executive Officer Mark Zuckerberg made Llama open source because he wanted to gain more control over the next significant technological wave, but this tactic works if Meta’s technology is widely used.
Last week, Zuckerberg stated during an earnings call that he wanted to convince the public sector to adopt Llama access to the US government. The most recent update may help to achieve that objective and open doors for Zuckerberg to develop relations with public officials.
Over the years, Meta was criticized for failing to protect young people online and its role in regulating user speech. In recent months, former President Donald Trump has made negative remarks about Zuckerberg, even suggesting that he might try and put Meta founder in jail if he gets re-elected to the White House.
Meta is also framing its choice to work with US defence agencies to prevent US enemies like China from gaining ground. According to Clegg, they are doing it while keeping in mind America’s best interests and the larger democratic world for American open-source models to flourish and outperform models from China and elsewhere.