How did US Military use Generative AI in intelligence operations? How did Pentagon collect data in sea off India?

The US is using generative AI tools in its espionage across the world. If media reports are to be believed, about 2,500 US service members from the 15th Marine Expeditionary Unit sailed aboard three ships throughout the Pacific Ocean. They conducted training exercises in the waters off India, Indonesia, the Philippines, and South Korea. In the course of their journey, the service members used the large language model, developed on the instructions of the Pentagon, and carried out tests using the generative Artificial Intelligence. 

Pentagon works on Generative AI tools

According to the reports, defense-tech company Vannevar Labs developed the generative AI tools at the cost of $99 million, fully paid by the Pentagon's startup-oriented Defense Innovation Unit. Established in 2019 by veterans of the CIA and US intelligence community, the tech firm develops artificial intelligence models that interpret data for warfare and surveillance. It proves how the Pentagon is harnessing AI to collect information and data and interpret them at a speed never imagined before. 

How cutting-edge technology is revolutionising the complicated world of espionage is clear by the fact that the officers onboard the ships used the new system to collect hundreds of thousands of pieces of open-source intelligence, non-classified articles, reports, images, and videos. These were collected in the various countries where they operated. What made the difference was the speed at which the data were interpreted and analysed. 

AI models used to interpret data

Talking to the MIT Technology Review, Captain Kristin Enzenauer, said that she had used large language models to translate and summarise foreign news sources. On the other hand, Captain Will Lowdon used AI to write the daily and weekly intelligence reports he provided to his commanders. He said, "We still need to validate the sources. But the unit’s commanders encouraged the use of large language models because they provide a lot more efficiency during a dynamic situation."

It was way back in 2019 when the Pentagon asked Vannevar to use large language models of the time to solve the data conundrum. It was found that the technology could enable AI not just to collect data but to analyse them like never before. Vannevar's AI tools were deployed in the Pacific. Enzenauer and Lowdon said that they carried out a double-check on the result of the AI’s work, and they didn't find inaccuracies.

Pentagon to spend $100 million on AI tools

The Pentagon is bullish on using AI tools and spending on the same. It announced it would spend $100 million in the next two years on developing generative AI applications. The US Defence Department also asked Microsoft and Palantir to work together on AI models that would make use of classified data.