CIA Builds Own Artificial Intelligence Tool – Report

US intelligence agency, the CIA, is reportedly building its own AI tool to assist in its intelligence gathering activities

The United States Central Intelligence Agency (CIA) is reportedly developing its own version of OpenAI’s ChatGPT-style tool.

Bloomberg reported that the US intelligence agency is developing the AI tool to help sift through the huge quantity of open-source information for clues. The tool will also reportedly analysts with their research.

The tech industry sometimes works closely with intelligence agencies. It is known for example that the NSA in the US and the the UK GCHQ have their own dedicated supercomputers to help in data mining and other operations.


AI tool

In October 2021 it emerged that Amazon Web Services (AWS) had won a contract thought to worth £500m to £1bn over the next decade, to host classified intelligence data.

Now according to the Bloomberg report, the CIA’s Open-Source Enterprise division plans to provide intelligence agencies with its AI tool soon.

There is no word if the CIA has brought in outside AI expertise, or whether it is working with any particular tech firm on the matter.

But the Bloomberg report stated that the CIA AI tool will used across the 18-agency US intelligence community, including those run by branches of the military.

“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,” Randy Nixon, director of the CIA’s Open-Source Enterprise division, reportedly told Bloomberg. “We have to find the needles in the needle field.”

The tool will allow analysts to see the source of information and ask questions about what they are viewing, Nixon reportedly said.

“Then you can take it to the next level and start chatting and asking questions of the machines to give you answers, also sourced,” he told Bloomberg.

Useful tool

The move comes just weeks after the CIA’s first chief technology officer, Nand Mulchandani, reportedly advised users to treat AI as “what I call the crazy drunk friend.”

Nand Mulchandani reportedly made the comments at the Billing Cybersecurity Summit in Washington.

Mulchandani said that while AI programs can provide useful answers for pushing analysts out of “conceptual blindness,” any information provided should be treated with appropriate suspicion.

He reportedly added the tech was “absolutely fantastic” when dealing with large amounts of data and looking for patterns.

Intelligence gathering in the online world is a regular occurrence nowadays, especially when considering offensive cyber operations carried out by hostile nations.

Earlier this year for example the ‘Five Eyes’ intelligence agencies, as well as Microsoft, warned that critical infrastructure in the US was being spied upon by state sponsored Chinese hackers.