The U.S. Central Intelligence Agency says it’s building its own ChatGPT-style chatbot to help analysts sift through enormous amounts of data so it can improve its investigations.

The CIA’s chatbot will leverage advanced generative artificial intelligence to help its agents access and understand open-source data, the agency said. The technology was developed by the CIA’s Open Source Enterprise unit, and will roll out the chatbot to multiple intelligence agencies besides the CIA, as part of an ongoing effort to rival the growing, AI-powered intelligence capabilities of China.

Randy Nixon, director of the CIA’s AI unit, told Bloomberg that the agency needs AI to keep up with the vast amounts of data at its disposal. “We’ve gone from newspapers and radio, to newspapers and television, to newspapers and cable television, to basic internet, to big data, and it just keeps going,” he said.

According to Nixon, the CIA and other agencies face a significant challenge in analyzing the level of data that exists across the internet. “We have to find needles in the needle field,” he explained. “The scale of how much we collect and what we collect has grown astronomically over the last 80-plus years. So much so, that this could be daunting and at times unusable for our consumers.”

Bloomberg said the U.S. military and intelligence apparatus is facing mounting pressure to compete with the rapidly advancing capabilities of China, which has publicly stated its ambition to become the world leader in AI by 2023. A telling sign of China’s ambitions, and its determination to make AI available to its law enforcement and intelligence officials, came in 2021 when it announced it had created a “prosecutor” that could press charges against criminal suspects with a 97% conviction rate.

But in the U.S., the government has been criticized over its sluggish adoption of AI for investigative purposes. Last May, National Security Agency Director Gilbert Herrera told Bloomberg that the intelligence community was still trying to identify ways it could use AI effectively, without violating privacy.

Nixon said the new chatbot being developed by the CIA will help to change the perception that the U.S. is lagging behind. It will enable investigators to scour the entire public web, condense that information and summarize it in a way that’s most helpful to them.

The tool will enable users to see the source of whatever information it provides, and, just like ChatGPT, they’ll be able to chat with it to receive more succinct answers about the information they’re interested in.

“Then you can take it to the next level and start chatting and asking questions of the machines to give you answers, also sourced,” Nixon said. “Our collection can just continue to grow and grow with no limitations other than how much things cost.”

Charles King of Pund-IT Inc. told SiliconANGLE that while there may be some public apprehension over the CIA’s plan, it is an entirely sensible one given that both friends and enemies of the U.S. are launching similar initiatives. What’s more, the CIA is going well beyond what most commercial generative AI platforms are doing by citing the source of its answers. “Not only is the effort leveraging the potential value of generative AI, but it is also designed to show or prove the responses it generates,” King said. “This is an approach that many of the public, commercial generative AI platforms would be wise to follow.”

Nixon assumed his role as the director of the CIA’s Open Source Enterprise division in January, and has been tasked with speeding up the development of open-source intelligence, Bloomberg reported.

According to Nixon, the CIA’s chatbot will be made available to other intelligence agencies, including the Federal Bureau of Investigation and the NSA. The U.S. military will also have access to the tool.

Nixon declined to say which AI model the CIA’s chatbot would use. Nor would he reveal anything about how it’s being trained or what measures it would take to protect privacy.

Photo: Markus Winkler/Unsplash