Why data is the hard part of military AI: Accenture defense AI lead on Project Maven and autonomous tanks

Accenture's Defense AI and Data Lead Bharat Patel Points Out That Data Is Never"AI-ready", the real bottleneck is using use cases to drive data quality. He cited Project Maven, launched in 2017, noting that initial large amounts of data lacked relevant targets, leading to poor model performance after launch and requiring continuous collection of valid data. He also stated that autonomous tanks are further away than imagined; the 2035 goal is for vehicles to understand environments and make decisions, but autonomous systems will still make errors, requiring human-in-the-loop verification.