Supporting Federal Agencies in Their Implementation of the Executive Order on AI – Part 1
The Office of Management and Budget’s (OMB) Memorandum on Artificial Intelligence (M-24-10) (the “Memo”), part of the President’s comprehensive strategy on AI, outlines a structured approach for federal agencies to leverage artificial intelligence responsibly. The Memo provides direction that establishes a governance framework, fosters responsible AI innovation, and implements important risk management protocols.
Microsoft values the Memo’s guidance that enables agencies to implement AI safely and effectively in fulfilling their missions, is committed to supporting these initiatives, and offers agencies the capabilities and solutions necessary for the seamless integration of AI across government functions.
Key Pillars
The Memo’s framework guidance can be broken down into four key pillars:
- Governance & Oversight: The Memo calls for the establishment of robust AI governance structures, including Chief AI Officers and AI Governance Boards, to ensure that AI deployments align with responsibility standards and the nation’s values. This foundational governance helps ensure that AI innovations serve the public good effectively and transparently.
- Responsible AI Innovation: Building upon solid governance, the Memo encourages agencies to drive responsible AI innovation and foster public trust. This commitment to responsible innovation enhances the government’s ability to deliver AI-powered services that genuinely benefit the public, while safeguarding individual rights and safety.
- Risk Management Practices: With innovation in place, the Memo emphasizes the need for comprehensive risk management practices. By integrating safety, transparency, and public engagement, these practices aim to mitigate potential harms so that AI technologies can be deployed responsibly and with minimal risk to public safety and rights.
- Compliance: Finally, the Memo encourages ongoing diligence in applying responsible AI practices, fostering a continuous improvement culture, and adaptation in AI governance. This proactive approach ensures that AI applications remain aligned with evolving standards and societal expectations.
Overall, the Memo necessitates a shift in how agencies deploy AI with the establishment of new guidelines. These guidelines are intended to ensure that the technological benefits are realized, while minimizing the risk of compromising public trust or safety.
Given the Memo’s December 2024 deadline, it is essential agencies begin work now to implement processes and enforce safeguards for AI systems that affect public safety or rights with additional measures for transparency, oversight, and community involvement.
This multi-part blog series provides practical guidance and tools from Microsoft to assist agencies in aligning to and implementation of the framework outlined in the Memo.
How Microsoft Can Help
Responsible AI Innovation
Microsoft remains committed to being a champion of Responsible AI. As the first major cloud provider to put forth guiding principles that prioritize fairness, reliability, safety, privacy, security, inclusiveness, transparency, and accountability in AI-enabled applications, Microsoft continues to operationalize responsible AI practices, tools, and governance throughout all AI products and services. We’re making these assets available to our government customers to do the same.
At the core of Microsoft’s responsible AI program is the Responsible AI Standard – a framework Microsoft uses to guide the responsible design and development of its own AI systems. Agencies can easily leverage the Responsible AI Standard as a guiding example in forming their own responsible AI policies and standards. To assist with implementation of the standard, Microsoft created the Responsible AI Toolbox – a collection of integrated tools and functionalities that can help operationalize responsible AI in practice. With the capabilities from the Toolbox, agencies can assess their models and make user-facing decisions faster and easier. Using the Responsible AI Dashboard, agencies can easily flow through different stages of model debugging and decision-making through a customizable experience.
Embrace Responsible AI Principles and Practices: Visit the Microsoft Learning Path
AI Maturity Framework
The Memo also calls on agencies to conduct a comprehensive AI maturity assessment to help understand where an agency stands today and the course necessary for achieving operational proficiency in the adoption of AI. The essence of this action-item is future-focused on preparing agencies to harness AI’s potential responsibly and effectively, while aligning to the principles of responsible innovation, governance, and risk management, as established by the President’s Executive Order on AI.
The following AI Maturity Model articulates the various stages of AI transformation for organizations. This model can help agencies understand where they are in their journey and core characteristics of organizations on their path to AI maturity.
The model guides organizations through four stages of the maturity process:
- Foundational
At this stage, organizations seek to understand the definition of AI and its applicability across various use cases and how others are utilizing AI in government to learn how to apply it. To foster AI adoption, organizations at this stage should invest in iterative experimentation, develop a data-driven culture and explore AI use cases. Out-of-the-box AI services such as Azure OpenAI can help agencies overcome operational challenges and transition more quickly into a digital organization.
- Approaching
Active implementation of cultural changes occurs in this stage with the purpose of empowering employees and enabling data-driven decisions. Investments at this stage center on understanding AI implementation, monitoring, and continuous improvement with rapid experimentation and establishment of AI governance protocols.
- Aspirational
Organizations at the aspirational stage understand that AI will be instrumental in helping them transform operations, make better decisions, and enhance mission delivery. Risk taking, transitioning away from sequential fixed projects to more iterative projects, and adoption of AI services to reduce operational complexity of maintaining core AI or to build custom models and mission-specialized Copilots becomes reality.
- Mature
At this stage the culture shifts to include lifelong learning with a growth mind-set. Rapid, iterative experiments come naturally. Operationalization of responsible AI processes; tools and practices; while maintaining AI talent; prioritizing new strategic initiatives; and continuing agile experimentation are required areas of focus for mature organizations.
Implementation of Microsoft’s AI Maturity Model – when coupled with our suite of cloud and our comprehensive AI platform – provides agencies with the capabilities to advance their AI maturity.
Enterprise capacity & IT Infrastructure for AI innovation
The journey towards transformative AI operations requires a thoughtfully considered strategy for innovation. While agencies have options on how they fulfill the Memo’s requirements, the adoption of a full-suite cloud infrastructure with an integrated, comprehensive AI platform offers a more straightforward, scalable solution.
The advancement of AI within federal agencies hinges on the establishment of a robust and agile IT infrastructure capable of supporting the entire AI lifecycle. Microsoft’s Azure Cloud infrastructure provides a purposefully designed environment to address these critical requirements, offering a comprehensive solution that enables agencies to harness the full potential of AI innovation responsibly and effectively.
Azure’s advanced AI infrastructure includes:
Comprehensive AI Platform: Azure provides agencies the flexibility to utilize AI Copilots out-of-the-box or infuse advanced AI services into applications using the latest industry leading models including Azure OpenAI, other powerful open-source models including Llama2, Mistral, and even leading-edge, small language models like Microsoft’s Phi-2 that can enable Generative AI in hybrid or edge environments. All of these models can be grounded and fine-tuned to the needs of the mission. Microsoft’s end-to-end AI platform also provides AI tooling and high-performance computing agencies can leverage to train custom-built models.
In addition to the comprehensive AI platform to support agencies’ adoption and innovation with AI, Microsoft Azure further supports agencies’ enterprise capacity and AI implementation with infrastructure that provides:
- Data Management and Governance: Advanced tools for data storage, sharing, and governance. Your data is your data.
- High Performance Computing: cost-effective, real-time, scalable compute resources essential for AI model training and inference tasks.
- Comprehensive Tooling and Libraries: Access to a wide range of software tools, open-source libraries, and development environments housed within our cloud.
- Deployment and Continuous Monitoring: Advanced monitoring and deployment tools ensuring optimal performance and quick identification and remediation of any issues.
- Cybersecurity and Compliance: Comprehensive cybersecurity and compliance measures, to ensure AI applications safeguard sensitive government data.
By providing a secure, scalable, and sophisticated environment for AI development and deployment with an integrated, comprehensive AI platform, Microsoft Azure is an agile environment in which to leverage AI technologies that adapt seamlessly to the evolving demands of AI applications and enhances AI maturity, all within a framework of responsible and secure IT practice.
In addition to Azure’s agile and scalable infrastructure, Microsoft also offers tailored Envisioning and Innovation workshops, providing agencies with a window into the best practices and latest trends in AI and equipping them with the insights needed to harness AI technologies effectively. These sessions support agencies in exploring the full potential of AI within a secure, governed framework.
For additional support in building capacity and the infrastructure needed for AI deployment and innovation, Microsoft has created a Cloud Adoption Framework to help get started.
By embracing a strategic approach to cloud adoption and harnessing the power of cloud computing to drive innovation, federal agencies can accelerate their AI initiatives, enhance service delivery, and fulfill their missions more effectively.
In turn, by simplifying the complexity of scaling the infrastructure needed for AI innovation, Microsoft helps agencies establish the necessary infrastructure to deploy AI solutions that are not only technologically advanced but also aligned with the principles of sound AI governance and risk management. Our Microsoft Federal team is ready to partner with agencies on this journey, ensuring that every step towards AI adoption and maturation is marked by strategic insight, operational excellence, and unwavering security.
Getting Started with AI
Enabling an AI-ready organization goes beyond technology – from readying your data estate to instituting a hypothesis-driven culture to maximize AI benefits. Commercial enterprises have been at the forefront of AI transformation, and in this three-part series from the Wall Street Journal, Microsoft and enterprise leaders share how they are leveraging AI with thoughts on how to plan accordingly: Next Gen Technology Shift, Lessons from Enterprise AI Pioneers, How Leaders Plan for AI Success
In Part Two of this blog series, join us as we discuss specific ways and tools Microsoft offers to support you as your agency begins to work through implementation of the framework outlined by the OMB in its Memo.
Our AI strategy directly supports national objectives, driving innovation, enhancing security, and boosting efficiency. Get the latest news and announcements on AI for government delivered to your inbox at https://aka.ms/AIforGovUpdates.
Jason Payne Chief Technology Officer, Strategic Missions and Technologies, Microsoft Federal
Kent Cunningham Chief Technology Officer, Federal Civilian Government, Strategic Missions and Technologies, Microsoft Federal
Light
Dark
0 comments