In October the White House released the long-anticipated Executive Order on Safe, Secure, and Trustworthy Artificial Intelligence. This EO provides federal agencies the reassurance needed to move forward with their respective AI journeys while sharing insight on addressing the associated risks.
One of the key components of the EO is recruiting and retaining an AI-trained workforce to ensure AI innovation and risk management are possible. While the EO acknowledges that the current government workforce is not adequate to support such innovation, funding initiatives for workforce training and education opportunities are being prioritized to decrease this shortage. OPM also provides an initial list of fundamental skills and competencies for AI-related positions to support recruitment.
To note, OPM shared a memo in December regarding direct hire authority for AI-specific roles. This DHA ties back to the AI EO and will assist agencies’ efforts to increase AI capabilities in the federal government, highlighting that AI-focused positions are a “critical hiring need.” This memo is another clear indicator that AI is a top priority for government as we begin 2024.
As outlined in the AI EO, the chief AI officer (CAIO) will coordinate, oversee and lead agency-specific AI initiatives, as well as spearhead internal policy decisions, regulations and workforce recruitment to implement and utilize AI tools effectively. As each agency implements the EO and reacts to the DHA, diligent decision processes are underway to provide agencies with the proper resources and leadership to take advantage of AI technology and its function.
Establishing a CAIO
CAIOs will be responsible for strategizing AI personnel recruitment, regulatory reporting, use case submissions and internal policy development – this position will be pivotal to each agency’s AI implementation success.
To quickly address the need for a CAIO, agencies may assign their chief technology officer (CTO) the role, but ultimately, the roles will be separated. Some may draw a parallel to the establishment of the chief data officer (CDO) role – prior to this position’s refinement, federal data management often fell into a gray area between the IT and operations departments. But in each AI office, hiring a technical AI lead will be imperative to ensure each AI model is properly managed and is producing the expected outcomes, and if not, corrections are being made based on feedback.
Similar to the establishment of a CDO, there is very little current guidance on what the CAIO role entails and it will take time for government to provide playbooks and other resources. While agencies work to refine this role, government contractors can fill the gaps with a chief AI office as a service (CAIOaaS) offering.
With their industry knowledge and perspective, contractors can help agencies through the first steps of establishing the CAIO, including office and strategy development, use case inventory and assessment, and governance and privacy compliance development. By engaging with industry to accelerate the first steps of establishing a CAIO, agencies can focus on their ethics, compliance education and performance, which will be the backbone of being able to implement AI.
For example, engaging with a company that works with existing COTS products already in enterprise architecture to explore AI, like ServiceNow, will serve as a better starting point as opposed to procuring AI specific tools. This approach can allow agencies to accelerate its demand and project management, manage internal inventory for regulatory check-ins and maintain agency CAIO support services in line with best practices and regulations established by government.
Engaging with a leading technology team to accomplish the initial steps will ensure compliance with relevant regulations and foster a culture of responsible AI development and deployment.
An effective AI strategy
Once the office and CAIO are in place, teams can begin to utilize AI technology to fulfill agency missions based on best practices.
Internal facing processes are the most straightforward and risk-averse area to begin integrating AI. For example, the IRS is utilizing AI in the form of robotic process automation (RPA) with software robots to improve internal procurement processes and modify contracts, reducing errors and administrative burden on employees. Through the internal deployment of this technology, the data used is controlled and remains highly confidential. Once this process has been practiced and revised, the IRS can deploy it publicly and apply it to taxpayer-facing use cases.
Other internal AI use cases include using generative AI for service desk support, specifically incorporating natural language processing (NLP) to map caller sentiments and pinpoint pain points more accurately.
A great example of NLP in action is at the Environmental Protection Agency, where models gather user feedback from surveys and sentiments regarding portal usage when seeking support. With this tool, employees can collect accurate data to ensure proper follow-up protocols, enhancing the customer experience and ensuring all citizen needs are met.
Another simple yet very effective use of AI internally is using RPA to automate mundane tasks such as document processing. Not only do these tools increase personnel productivity, but they allow employees to focus on higher-value tasks, increasing their worth as an employee.
Since 2017, government spending on AI-related technologies has increased by nearly 2.5 times – there is a significant opportunity within government to capitalize on this technology to improve internal processes, outcomes and citizen experience.
Under the guidance of the CAIO, agencies will be able to understand when and where AI services can safely be deployed. By using AI within internal processes first, there are lower associated risks and employees can address issues before deploying AI tools to the public. Once AI is offered to the public, the risk becomes much higher as citizen data is involved.
The establishment of CAIOs across federal agencies will play a pivotal role in AI success across government and is an excellent opportunity to engage with industry partners to accelerate adoption.
Laura Stash is Executive Vice President of Solutions Architecture at iTech AG, an Arlington, Virginia-based technology consulting firm.
