This article is an excerpt from GovLoop’s recent guide, “Emerging Tech in Government: What It Means for Your Career.” Download the full guide here.
An interview with Dave Jones, Product Marketing Director, Nuxeo
Advancements in digital capabilities are transforming the way agencies create and manage data, whether it’s content from social media platforms, law enforcement surveillance footage or mobile imagery.
Given the volume of information that agencies gather on a daily basis, they need modern tools to help them effectively manage and make sense of their data. To keep pace with these growing demands, agencies are increasingly supplementing human intelligence with artificial intelligence (AI).
To explore how AI is enhancing content management and data usage in government, GovLoop sat down with Dave Jones, Product Marketing Director at Nuxeo, which specializes in making enterprise content services accessible, flexible, scalable and AI-enabled.
“AI is starting to become heavily used in government,” Jones said. “But probably 90 percent of that use is behind the scenes and is focused on automation.”
Consider the information collected from drones, traffic cameras and citizen-facing portals that agencies gather, manage and analyze. That information is of no use to government if it’s sitting dormant in multiple systems or databases. This is where AI comes into play.
“It might be as simple as scanning through piles of invoices every month and processing them more efficiently than humans,” Jones said of current AI applications. “It could be producing thousands of automated emails, letters, text messages, or informing those with benefits claims of their entitlements. Or it could be as simple as providing scheduled roadwork updates through online services such as Google Maps.”
Jones also highlighted more advanced use cases for AI. For example, imagine a homeland security system that could trawl through live surveillance video footage to identify the location and activities of known criminals. Or a welfare system that could automatically identify and provide assistance to citizens struggling with mental health issues based on their social media posts. The potential benefits in data classification are vast and far-reaching.
Essentially, these AI applications are trained using a sample set of data, learning how to extract key information from whatever document or systems they trawl. That allows the applications to process large quantities of documents with minimal human interaction, which is a crucial function considering the expanding size and scope of data sources.
“We’ve currently been hit with information overload,” Jones said. “If you think about the amount of information that you receive personally on a daily basis, scale that out to an agency and you start to see the challenge.”
Of course, there are concerns regarding AI, including the fear that it will replace employees and threaten their jobs. Jones understands the concerns, however, he believes that’s not the case.
“The purpose of AI is not to replace humans, but to make those humans more engaged with the job that they’re doing and also to cut the dreariness out of daily tasks,” he said. “It’s meant to actually get the right people doing the things that they’re skilled in and paid to do.”
Additionally, machines don’t tire of their workloads. They can complete tasks with a higher degree of accuracy and for a longer period of time than their human counterparts.
“Humans tend to, by nature, take shortcuts; AI doesn’t,” Jones said. Likewise, human performance tends to decrease as a project increases in size. But AI delivers the same level of accuracy every time you use it. In fact, it gets better the more it is used.
Due to its ability to complete repetitive tasks with record accuracy, modern content services platforms are utilizing AI to help agencies better classify and manage their data. Take the earlier use case of a homeland security system trawling through surveillance video footage to identify criminals. Using AI tools law enforcement agencies can aggregate detailed identifiers from the footage like a hat, license plate or specific location. AI classification typically returns more than 20 tags per image, compared to 5 for humans, with that number decreasing the more images you ask them to classify.
“AI allows agencies to classify content much faster, and on a much wider scale. That allows agencies to make use of more data, faster and more efficiently than ever before,” Jones said.
It’s become clear that AI will factor heavily into how government operates in the coming decades. In terms of data processing and classification, the future will be automated, and government agencies of all sizes must be ready for it.