Software

4 Ways Dell Technologies Is Helping Agencies Establish Generative AI

Generative artificial intelligence will help simplify AI infrastructure management, strengthen cybersecurity, enhance predictive modeling and more.

Brian T. Horowitz

Brian T. Horowitz is a seasoned tech journalist and copy editor with 25 years of experience.

As agencies prepare to increase adoption of generative artificial intelligence, they will need to simplify how they source, deploy and manage AI infrastructure.

While evaluating generative AI’s many possibilities, agencies will likely consider a mix of cloud, edge and on-premises infrastructure. Cloud-based services for AI include Dell’s APEX for generative AI, which improves agencies’ ability to scale computing power based on real-time data use, says Paul Perez, technology senior fellow for Dell Federal at Dell Technologies.

“It also lowers the barrier to entry by making AI-optimized infrastructure more accessible, driving faster time to value,” he says.

Improving mission effectiveness and lowering the cost of services are some potential use cases for generative AI, he adds.

The Department of Energy’s Office of Science could turn to generative AI to increase the rate at which new galaxies are discovered throughout the universe. Other agencies have announced plans for space applications, nuclear reactor simulations and event correlations. Generative AI could also help agencies predict weather, prevent wildfires and manage critical infrastructure, Perez says.

Agencies are already using or planning to use generative AI to improve the following four areas.

EXPLORE: Modernize your data center with Dell Technologies solutions.

Simplifying AI Sourcing with a Structured Approach

As agencies prepare to deploy generative AI, they can simplify the sourcing by following a structured approach, Perez says.

Consider the volume of data, computational needs and budget constraints when defining AI requirements. Conduct market research to evaluate AI-enhanced infrastructure solutions, including cloud platforms or on-premises hardware.

As agencies evaluate generative AI possibilities, they will work with a large number of documents as knowledge bases rather than large amounts of data, Perez says.

“The most important thing is to realize that all AI/generative AI use cases don’t require large farms of power-hungry eight-GPU servers,” Perez says. “For example, we can help our federal customers rightsize computing and storage infrastructure depending on what model is being used, phase of life cycle — training, tuning, inference — and size of the model.”

Services such as Dell Technologies Validated Designs may lower the energy and deployment costs of AI infrastructure and speed up the time it takes to achieve desired outcomes, Perez says. That’s particularly relevant “when we consider that the bulk of data in this world is now created outside of data centers at what we call the edge, and data is hard to move.”

Click on the banner to learn how platform engineering works.

ft-platformengineering-animated-2024-cta-desktop

ft-platformengineering-animated-2024-cta-mobile

Predictive Modeling and Scenario Planning

Agencies can plan for future events more effectively when using generative AI in predictive modeling scenarios. Generative AI provides access to high-quality data and the necessary computing resources for such exercises. Agencies can also perform scenario planning to better understand how generative AI can help by developing protocols on how to incorporate generative AI into planning that’s already taking place, Perez says.

Identify key decision points where generative AI can provide valuable insights and ensure that the tools are used in conjunction with other analysis to enhance or inform, rather than replace, human judgment, he says.

Strengthening Cybersecurity Measures and Protocols

One of the main areas agencies now use or intend to use generative AI is in cybersecurity.

“Generative AI can be used for anomaly and threat detection, allowing agencies to improve their ability to detect and respond to cyberthreats in real time,” Perez says. “This complex-event correlation capability also enables better management of critical national infrastructure.”

As agencies prepare for generative AI, they will require new software to enforce data governance policies, data privacy and security, as well as access controls and data grooming, Perez notes. Security and privacy software will allow agencies to guard against unauthorized access and achieve compliance, he says.

“In certain mission-critical use cases, fault-tolerant methods such as Byzantine consensus may be necessary to secure generative AI/AI environments against malicious attacks,” he adds.

Transforming Employment and Citizen Engagement

Some agencies will upgrade and improve employment and citizen engagement experience using AI in health services or duty services, speeding up response times to citizen inquiries, Perez says.

For example, agencies can use AI to improve lookup times when finding information for citizens or responding to requests. A generative AI copilot could provide real-time guidance to federal employees as they provide essential information to citizens, Perez predicts.

The federal workforce “spends significant amounts of time looking up policies for guidance,” Perez says. “This lookup time is experienced as wait time by the colleagues or citizens they are trying to serve. Imagine instead how a generative AI copilot trained on all relevant policies and guidelines and receptive to the conversation or the IM exchange could offer real-time guidance to employees.”

UP NEXT: Agencies should be a part of the AI proof-of-concept process.

Going forward, an entirely new set of AI-enabled software tools will provide agencies with everything they need to perform key tasks such as data collection, Perez says.

“A comprehensive suite of software tools that support the entire lifecycle of generative AI projects — from data collection and model training to analysis and decision-making — will be essential as agencies look to leverage the full potential of generative AI insights,” Perez says.

Brought to you by: