The Army is launching a new pilot to support a broader push toward the service’s adoption of generative artificial intelligence, officials announced Friday.
The effort, called #CalibrateAI, follows other efforts — such as #DefendAI, #BreakAI and #CounterAI — that have been launched as the agency pursues a 500-day implementation plan to adopt these types of technologies and overcome some of the issues. the challenges that come with it.
“Gen AI models offer unique and exciting opportunities for the military. These models have the potential to transform mission processes by automating and executing certain tasks with unprecedented speed and efficiency. Commanders and senior leaders should encourage the use of Gen AI tools for the right use cases,” Army CIO Leonel Garciga said in a directive issued to the department in June.
Generative AI systems can create new content – such as text, images, audio and video – based on the data they are trained on and human cues. ChatGPT, a popular large-language model, is a prominent example.
While much attention is being paid within the defense community to how artificial intelligence could support and enable U.S. military operations on the battlefield, officials are also eager to deploy the technology for back-office functions and take on some of the drudgery that humans previously had to do. have done. had to perform.
On Friday, Jennifer Swanson, deputy assistant secretary of the Army for data, engineering and software, announced the launch of #CalibrateAI, which will explore “innovative applications” of the technology for Army acquisition activities.
According to officials, the initiative will use an “advanced” capability developed by the nonprofit LMI at no cost to the military.
The tool, which aims to “simplify repetitive and time-consuming tasks,” will use data analytics, machine learning and natural language processing to “deliver tailored answers that are highly relevant to specific topics, increasing efficiency and effectiveness of information retrieval and analysis is improved,” the Army said in a news release.
However, officials are also concerned about ensuring the security and accuracy of Defense Department data and the output of gene AI tools.
In his June memo, Garciga noted that these technologies present unique challenges related to data privacy, security and control over the content generated. “Therefore, its use should be carefully evaluated and monitored,” he wrote.
The #CalibrateAI capability will be deployed in a secure Impact Level 5 (IL5) cloud environment and will be able to process Controlled Unclassified Information (CUI) data. It includes customizable user access controls to protect need-to-know information, according to the release.
The military is concerned about possible hallucinations, a trend seen in other generative artificial intelligence systems launched in the commercial sector. For example, large language models sometimes produce responses to user input that are factually incorrect or otherwise problematic.
Recognizing these potential pitfalls, the new pilot will also focus on “identifying and flagging potential ‘hallucinations’ or erroneous results, thereby increasing the reliability of AI-generated content,” the release said.
“By leveraging off-the-shelf AI tools and leveraging reciprocity between services provided by DoD CIO, #CalibrateAI will explore how we can increase productivity while improving information accuracy,” said Swanson in a statement. “The ability to query curated document sets to generate new content, along with providing citations, will ensure that our results are not only accurate but also easily fact-checked.”
The pilot’s broad objectives for the acquisition community include increasing productivity, improving accuracy, fostering innovation and identifying a cost-effective route for broader adoption of the technology. According to officials, this will be achieved through the use of AI tools to collect, manage and generate critical information relevant to acquisition activities, implement mechanisms to provide citations and promote the exploration of new applications of artificial intelligence in acquisition.
