The ability to shift processing and analytics power directly to disconnected, remote locations is a game changer for public sector operations.
JUMP TO...
AI at the edge enhances
decision-making
in unpredictable environments
Challenges to AI
at the edge
Delivering AI capabilities
Use cases for AI
at the edge
As the public sector progresses in its artificial intelligence (AI) journey, it’s increasingly seeking ways to implement AI-powered solutions throughout government operations, from fully equipped office environments to the tactical edge. Bringing the equipment and compute power required to successfully deploy AI in remote, complex environments, however, introduces additional hurdles.
“Significant efforts are underway to integrate AI into all aspects of government at the enterprise level. One of the most challenging questions is how do we deploy AI capabilities to the places it is most needed, especially in times of crisis?” said Matt Marsden, Head of Customer Engineering for Defense at Google Public Sector. “How can we deliver the same global-scale capabilities available to the enterprise to the disconnected, contested environments where it is needed most?”
Delivering AI capabilities to the edge largely depends on overcoming connectivity challenges. The logistical concerns that arise in remote environments start with being able to tap into the bandwidth and infrastructure support necessary to run sophisticated AI workloads.
“In the office, most people have high bandwidth and stable connectivity that facilitate access to hyperscale resources,” Marsden said. “As you move to the tactical edge, it becomes far more difficult to access mission critical resources.”
The concept of “the edge” is much more expansive than the tactical edge, and challenging environments aren’t limited to defense or national security. Even down to a local level, a natural disaster like a flood or a fire can destroy critical infrastructure and disrupt the ability to communicate in an emergency.
On top of basic logistics, given the sensitivity of government workloads, robust security is critical to leveraging AI capabilities in a disrupted or disconnected environment. Many solutions involve sending data back and forth to a central office, headquarters or data warehouse, which can introduce additional latency or vulnerabilities, impacting data security, integrity and, ultimately, decision-making.
“We know that we can make a dramatic impact on critical operations by pushing advanced analytics and AI capabilities to the edge, where they are needed most,” Marsden said. “Delivering capabilities for advanced data analysis in the field allows for more informed decision making.”
Challenges to AI at the edge
Delivering AI capabilities
The ability to provide compute, storage and processing capabilities at the edge has existed for some time now. The real game changer is enabling a fully capable cloud environment that lives directly at the edge. Google Distributed Cloud (GDC) enables users to build and train AI models on Google Cloud and deploy and run them locally at the edge.
GDC can essentially be a “cloud in a box” as Marsden described. In fact, it comes in a portable option. The GDC air-gapped appliance can be lifted and moved by two people to wherever it’s needed. It’s part of GDC air-gapped, which does not require any connectivity. It is one of three GDC solutions that extend Google Cloud to locations outside Google Cloud’s data centers:
Google Distributed Cloud (GDC) connected brings Google Cloud infrastructure and services closer to where data is being generated and consumed.
Google Distributed Cloud (GDC) air-gapped lets you host, control, and manage infrastructure and services directly on your premises.
Google Distributed Cloud (GDC) software-only enables you to run Google Cloud on your own hardware: “GDC connected delivers the AI processing and storage capabilities at the edge but sends data back to Google Cloud that allows for continuous operations during intermittent, disconnected states,” Marsden said. “Then there's Google Distributed Cloud air-gapped that is meant to be completely disconnected and have no requirement for connectivity back of any kind. This gives users the ability to build and deploy cloud workloads, and tap into critical services, in a completely isolated environment.”
Use cases for AI at the edge
From emergencies and natural disasters to day-to-day work in complex and austere environments, having a cloud-in-a-box solution available has the potential to revolutionize operations across government.
Real-time translation: When language barriers arise in the field, AI-enabled translation can be a lifeline, particularly in an emergency. “Being able to speak in the native language of the people that we’re supporting is critical. When you have translation capabilities hosted and supported by large language models and artificial intelligence capabilities in a product like Google Distributed Cloud, it gives us the ability to handle natural communication with people in times of crisis,” Marsden said.
Document processing: This entails the ability to scan or ingest documents in numerous formats — text, pictures, diagrams and more — into an AI application that can then analyze the documents and help make informed decisions about the content. GDC can complete this work right at the edge, whether to help analyze and understand local policy or doctrine, process a mass amount of insurance claims in a disaster, or translate documents — all tasks that would take humans alone significantly longer to perform.
Generative AI search: Consider someone suddenly deployed to another country on behalf of the state department, DoD, or for national security purposes. They’re expected to get up to speed on the local environment to make informed decisions as quickly as possible. Through Retrieval-Augmented Generation (RAG), they can upload a corpus of essential data and then query it to generate information about local people, customs, processes and procedures. “You can reference your responses in the data that you have uploaded, so all of your responses are grounded,” Marsden said. “With very little existing knowledge of the location or the situation, you have the ability to generate communications or protocol documents to rapidly understand how to work within the environment to which you’ve been sent.”
GDC fits seamlessly into the overall Google Cloud ecosystem. Its edge capabilities are protected by Google Cloud’s best security practices and built to meet stringent government security standards. This includes both the security controls of the GDC instance as well as the security controls of the individual applications deployed within that instance. In other words, users have the same role-based access controls, authorization and authentication, and encryption capabilities expected from any cloud environment available in a fully disconnected state.
“It’s about bringing the power of Google Cloud to the people that need it, at the time that they need it to enable informed decision making and data analytics at the edge,” Marsden said, “in a way that is easy to maintain and meet all of the security and compliance requirements of
the user.”
Learn more about the advanced AI capabilities Google Distributed Cloud can bring right to the edge.
The defense landscape extends to all ends of the earth and skies. Bringing analytics and processing capabilities directly to decision makers, even at sea, in the air or underwater, results in the most accurate, efficient and secure actions in critical moments.
Alongside the traditional Google Cloud Platform, GDC is an integral piece of the overall Google ecosystem. Learn more about
the Googlesphere.
