The Department of Defense’s new AI memo highlights the importance of ensuring that AI technology is reliable, governable and traceable, Dhanda wrote.
He noted that the industry has initiated steps to improve the interpretability of machine learning models and that there are tools and methods available for users to identify biased data and come up with documentation to explain how an ML model works.
“By enhancing operational transparency and oversight, agencies can ensure that they’re using ethical AI,” Dhanda said.
He said H2O.ai helps facilitate AI adoption for agencies by offering a platform designed to provide transparency, mitigate cyberthreats and other risks and optimize operations and by managing large data volumes using NVIDIA’s graphic processing units.
Dhanda also mentioned the company’s H2O AI AppStore platform for sharing ML models and how it could help agencies leverage AI.