The Air Force Bought a Surveillance-Focused AI Chatbot

The U.S. Air Force paid for a test version of an AI-powered chatbot to assist in intelligence and surveillance tasks as part of a $1.2 million deal, according to internal Air Force documents obtained by 404 Media. 

The news provides more insight into what military agencies are currently exploring using AI for, and comes as more AI companies eye the military space as a business opportunity. OpenAI, for instance, quietly removed language that expressly prohibited its technology for military purposes in January.

“Edge Al Platform for Space and Unmanned Aerial Imagery Intelligence,” a section of one of the documents reads. The contract is between the Air Force and a company called Misram LLC, which also operates under the name Spectronn.

Included in a “milestone schedule” explaining the specifics of the deal are the items “ISR chatbot design” and “ISR chatbot software.” ISR refers to intelligence, surveillance, and reconnaissance, a common military term. Other items in the schedule include “data ingestion tool” and “data visualization tool.” 404 Media obtained the documents through a Freedom of Information Act (FOIA) request with the Air Force.

On its website, Spectronn advertises an “AI Digital Assistant for Analytics.” It says the bot can take data such as images and videos, and then answer plain English questions about that information. “Current analytics dashboard solutions are complex and not human-friendly. It leads to severe latency (from hours to days), cognitive load on the data analyst, false alarms, and frustrated decision makers or end-users,” it reads. 

The documents also specifically mention a “ML [machine learning] model of SAR ship imagery analytics.” The rest of the documents are heavily redacted, but a demo video from Spectronn online shows what an analyst might do with such a technology.

A SCREENSHOT FROM THE DOCUMENTS.

In that demo, a user views an aerial image of two docked ships. The tool also provides an annotated copy of the image, with rectangles drawn around objects it seems to have detected. A “Detected Events” summary then reads “38 cars, 12 containers and 2 large ships are seen. Car traffic has increased by 30% over the last 2 weeks. The two ships have been docked for 1 week.”

A “Detected Objects” table then lists the various items, along with a confidence rating, such as a 0.720 confidence rating that a highlighted object is a ship, or a 0.881 rating related to a container. 

Clicking on the lower right-hand corner of the screen opens up the chatbot. “Hi, How may I help you?” it asks.

“How many ships were docked in the last week?” the user then types. The chatbot then provides a line graph allegedly showing the number of ships docked over time.

Next, the user asks the bot to “Show me the pictures of ships seen today?” The bot then presents two cropped images of the ships. Finally, the user asks the bot to again summarize the objects that have been seen today.

Bryan Ripple, team lead of media operations at Air Force Research Laboratory Public Affairs, told 404 Media in a statement that “The Department of the Air Force isn’t using this program at all, nor do we have any intention of using it for targeting decisions. The program is merely at the early stages of development to determine if there is a use case for the Department of the Air Force.” 

Rajarathnam Chandramouli, who is listed in the documents as Misram’s CEO and who also describes himself as the CEO of Spectronn on his LinkedIn, did not respond to a request for comment. 

A month ago, Chandramouli wrote in a LinkedIn post that his team had been awarded a “contract to further advance and demonstrate our trusted AI capabilities for Department of Air Force customers.” 

Spectronn’s site suggests other use cases of the technology, including “retail crime, pattern-of-life changes, cybersecurity attack.” The Spectronn website also includes a video testimonial from an apparent U.S. Space Force member, who says “utilizing AI and machine learning technology from innovative companies like Spectronn, could not only help us improve our speed and accuracy of our efforts, but reduce the demand on our operators.” 

Spectronn also advertises what it describes as “‘Alexa’ for Video Analytics.” The tool appears similar to the previously mentioned chatbot, but in this case is operating over WhatsApp and is allegedly monitoring feeds from cameras.

“I am monitoring your 4 cameras,” one message from the bot reads in a YouTube demonstration video. “How can I help you today?” Once asked “What happened yesterday in my backyard?” the bot says “there was a party from 7pm to 9pm.”

“I see a young couple having wine and fun by the firepit,” it adds before showing a photo of that couple. The people are highlighted with rectangles saying “person,” and objects around them such as “bottle” and “chair” are highlighted. It then shares a short clip of the particular event.