The U.S. Defense Department has a new office dedicated to advancing the use of AI throughout the military. Now the money from a $249 million blanket purchase agreement is starting to flow to several AI tech providers that also work with enterprises in other industries, including companies like Arthur that help monitor AI models to avoid bias and inaccuracy.
Arthur’s software tracks performance of AI models including computer vision and natural-language processing models. Arthur is among a burgeoning sector of companies providing software and services for auditing how machine-learning and AI models are built, and monitoring them for problems after they’re in use. The company collected $15 million in series A funding in December 2020.
The DoD also signed on earlier this month to use DataRobot’s cloud AI platform and services, including to detect, measure and prevent bias when building and operating machine-learning models.
Companies working with the Pentagon to provide AI and related tech have drawn their fair share of scrutiny, perhaps most famously when Google employees protested the company’s Project Maven effort to develop drone AI technology with the military. Less media attention went to C3.ai when it signed a massive $500 million deal to provide its enterprise AI products to all Defense Department agencies.
Defenders of the military’s push toward using AI point to ethical AI principles the Pentagon established in 2020, which included goals for responsibility and equity.