Government and private sector organizations that are seeking to create successful artificial intelligence (AI) technology efforts need to prioritize investments in data maintenance and categorization, two AI experts said this week.
Thomas Sasala, Chief Data Officer for the Department of the U.S. Navy, explained the importance of maintenance and categorization at ATARC’s AI and Visual Data Summit on June 28.
“You have to set the foundation, there’s a lot of pragmatic things you need to do with your infrastructure and data before you can even consider using artificial intelligence” said Sasala, who has served as Navy CDO since 2019.
“If you are investing in artificial intelligence, and you’re not investing in things like cleaning and curating and cataloging your data, and knowing what your data landscape looks like, then you are fundamentally doing it wrong,” he said.
Nick Psaki, Principal Technologist, Office of the CTO, at Pure Storage, reinforced the data-related aspects of that sentiment.
“Ai just works with data at a degree of parallelism, speed and scale that really requires a dedicated purpose-built type of infrastructure,” said Psaki.
“Those tools are adaptations of existing technology, the integration of those tools to create the capacity to be agile, to readily adapt to new data types to new questions to models, etc.,” he said. “There’s no substitute though for great data and people who know how to leverage AI algorithms to be used as a service.”
Both Sasala and Psaki also agreed that AI is no longer a technology of the future. Instead, they said, the technology is here now, and needs to be understood in that context.