By Eric Velte, Chief Technology Officer, ASRC Federal

As published in Forbes on November 14, 2023.

The application of emerging technology necessitates new policy frameworks and guidance. And with new policies constantly on the horizon, there is a race to understand the proper use and implementation of cutting-edge technology.

To implement this technology effectively, we need to have a firm grasp on the key constraints, risks and benefits.

For example, with the Pentagon’s Chief Digital and Artificial Intelligence Office (CDAO) teed up to release a unified Data Analytics and AI Adoption Strategy this year, it is crucial to understand the fundamental challenges organizations face regarding AI and ML. This strategy will guide the department in developing and adopting data analytics and AI capabilities.

As agencies and their industry partners thoroughly evaluate AI and ML challenges, they can move forward in parallel to secure dedicated AI-literate personnel, design training and lean on the contracting community for expertise. While this process can be accelerated, it cannot be bypassed.

Core Roadblocks to Federal Adoption and Implementation
Technological challenges and storage represent the two pivotal impediments obstructing AI/ML utilization in the federal landscape.

It may seem ironic, but technology itself poses the most significant challenge to advancing the federal adoption and implementation of AI and ML. Agencies face resource barriers in understanding what the technology can accomplish and how it can be used.

When it comes to embracing modernized technologies, the US Government is a heterogeneous mix of both early adopters developing cutting edge capabilities (for example the DOD S&T community), and more deliberate adopters who are slowly adopting AI/ML technology due to their heightened social responsibility to the public, and sometimes bureaucratic processes. Mony of the challenges faced by all parties stem from a need to understand if these large language models are appropriate for governmental operations and their ethical implications.

Another central technological challenge is the issue of privacy. Public trust is a necessity for government agencies. Due to the vast amounts of sensitive and personal data that agencies work with, early adoption can raise security concerns. Due to these privacy concerns, agencies typically take a wait-and-see approach before diving into new territory. From a data security aspect, agencies must be extraordinarily cautious when utilizing new tools, and that can limit AI uptake.

An additional roadblock to AI and ML model adoption and implementation is that they consume substantial amounts of data, resulting in the need for robust storage solutions. Due to the vast amounts of data AI models require to be trained and used accurately, this presents controversy around data rights as well as access, and storage challenges. Additional storage considerations include the need to store susceptible data securely and how to prompt AI. As a result, diverse models increase this burden geometrically. With the storage needed for these applications, the cost of maintaining and upgrading infrastructure can cause hurdles.

Cultivating a Conducive Environment to Adoption and Implementation
Identifying and understanding federal agencies’ limitations in adopting and implementing new technologies will be instrumental in addressing them.

To combat the need for knowledge on the proper usage of AI and ML, having dedicated personnel entrenched in these technologies’ inner workings is a way to encourage ethical use and lessen existing knowledge barriers. Enter the AI Operator: A dedicated role supporting the proper use of AI that will help the government’s speed in adopting AI innovation. This role will not only help bridge knowledge gaps but also support the concern for transparency and ethical use.

Developing training programs and skilling employees will also provide operators with the most up-to-date knowledge of AI and enhance decision-making. Providing ongoing support to ensure that the workforce is keeping pace with the digital era will be incredibly beneficial to encouraging the use of new technology. Furthermore, establishing clear directives for using emerging technologies and how they will play into day-to-day operations will strengthen implementation. Policies and regulations will support mitigating risk by directing proper usage.

Lean on the Contracting Community
A key factor for government agencies to consider is that the private sector under the right circumstances and acquisition strategies is willing to assume more risk than the federal government is able to, serving to shield agencies from crises and protect the public interest. Not only are government contractors able to help mitigate risk, they are also deeply familiar with the federal landscape. The contracting community has many thought leaders who can validate the proper usage and application of AI and ML and strengthen the federal government’s willingness to participate in adopting emerging technologies. A lot of companies like ours are willing to partner with the government transparently to examine the risks and benefits of emerging AI technology. The private sector as a whole is typically among the early adopters of next-generation technology, making it an ideal partner. Along with their ability to provide rapid response and be nimble, contractors are proven to be cost-effective and scale rapidly. Contractors can efficiently adapt to project needs and bring specialized expertise, bridging the knowledge gap. Government contractors provide an untapped potential and should be considered when looking for implementation support for new technologies.

Taking a multifaceted approach that considers fundamental governmental limitations will be critical to ongoing efforts to embrace transformative technology.

Looking to the future, there are several actionable steps that organizations can take further to advance the adoption and implementation of AI and ML. Showcasing industry thought leadership will support validating the proper usage and application of AI and ML, strengthening the federal government’s willingness to participate in adopting emerging technologies. Additionally, once these technologies are appropriately vetted and frameworks are in place to guide ethical usage, agencies will be more inclined to adopt and optimize AI and ML across the federal landscape.