Skip to content

Microsoft's DoD Azure Summit: JADC2 Data Fabric, PKIs, and Fortifying U.S. Decision Advantage

Attended a military conference featuring the newest Azure mission capabilities, such as GPT-4 Copilot Semantic Kernel, Hyperscale AI, Ask Sage, and more.

Kyle Lyon
Kyle Lyon
7 min read
Microsoft's DoD Azure Summit: JADC2 Data Fabric, PKIs, and Fortifying U.S. Decision Advantage
DoD Azure Summit - DTS

Table of Contents

Introduction

The explosive evolution of artificial intelligence, showcased by ChatGPT and Bard, has undoubtedly stirred the defense industry. This fact was brought to life during the recent Department of Defense Azure Summit hosted by Microsoft.

In May, I had the pleasure of attending this event that brought together cloud engineers, decision-makers, and operators from the DoD. The summit served as a platform for sharing knowledge and ideas, revealing the immense capabilities of large language models (LLMs) and the transformative potential of military digitization.

The event featured informative sessions grouped into specific learning tracks. The summit's highlight was the tour of Microsoft's digital showcase facility. I was particularly impressed with the interactive digital experiences explicitly designed for national security missions.

Microsoft's Digital Showcase Facility

Several industry leaders delivered insightful keynotes, providing perspectives across different solution sets and mission ownership. A focus was placed on the latest innovations from Microsoft (of course), particularly Azure, and the need for cloud capabilities in the military.

Microsoft Cloud for Mission

During the conference, several innovations captured my attention due to their potential to disrupt military operations significantly. Here are the most noteworthy ones:

  • The partnership of Microsoft's Co-Pilot and its Planetary Computer.
  • The blend of GPT-4 and the Semantic Kernel marrying LLMs with traditional programming languages, such as C# and Python.
  • "Ask Sage" and its capacity to generate mission-specific responses, especially when contrasted with the capabilities of ChatGPT, is profoundly impressive.

In the following sections, I'll discuss each game-changing technology and its potential to transform warfare.


Co-Pilot & The Planetary Computer

Microsoft's Planetary Computer combines a multi-petabyte catalog of global environmental data with APIs, creating an environment to provide up-to-date information about the Earth's ecosystems. I think it was initially intended to promote conservation and sustainability efforts on a global scale.

Here are a few examples of analyses you can do with the Planetary Computer shown below.

From a military perspective, the Microsoft Planetary Computer could offer several benefits:

  1. Environmental Intelligence: Global ecosystem data can aid military planning by providing insight into terrain and weather conditions.
  2. Disaster Response: It can aid disaster response planning and execution with real-time environmental monitoring data (paired with predictive models) during natural disasters and climate crises.
  3. Security and Defense: Satellite imagery (and predictive modeling) can aid security and defense, such as monitoring border areas and conflict zones.
  4. Infrastructure Planning: Environmental data can inform where to build military infrastructure less susceptible to environmental risk and less vulnerable to climate change.

In the commercial sector, analysts often complain about the lack of data. However, the defense industry faces the opposite problem with overwhelming data. Microsoft is addressing this issue by introducing a natural language prompt to the Planetary Computer and enabling Co-Pilot to identify relevant data and perform analysis on behalf of the user. This allows the user to interact with the data and receive visualizations in response. Co-Pilot significantly improves an analyst's ability to process data by 100x, and since Co-Pilot works with actual data, it cannot hallucinate like it does when generating text. However, it may make programmatic errors, which the analyst must verify and confirm the logic used by the computer to reach a solution.


GPT-4 & Semantic Kernel

By now, you probably are already aware of GPT-4, but for those who are not, GPT-4 is a tool developed by OpenAI that can do many things, like processing language and generating text. Microsoft also has a lightweight SDK called Semantic Kernel to connect GPT-4 with common programming languages, such as C# or Python.

This means developers can integrate GPT-4 with SK using a specialized API to call upon GPT-4 within their code. This API allows for a variety of features that simplify the integration of GPT-4 into applications, such as:

  • The ability to pass in natural language prompts GPT-4.
  • The ability to receive natural language responses from GPT-4.
  • The ability to control the behavior of GPT-4 through parameters.

This means military programmers and developers can now leverage the capabilities of GPT-4 and SK working in tandem, opening up many possibilities for various applications, some of which I have listed below:

  1. Data Synthesis & Debriefing

    GPT-4 can help to synthesize vast amounts of data and information, such as intelligence reports, satellite imagery data, weather conditions, troop movements, etc. By integrating this with SK, the GPT-4 could provide concise and relevant debriefings to military command.

    So after-action reports could be automatically generated and summarized immediately after an event. GPT-4 could parse through large amounts of data, pull out the key points, and present them concisely and understandable. I think this would usually take weeks to perform or even months when everything used to be done on paper.
  2. Multimodal Question-Answering

    GPT-4's capabilities to understand and respond to natural language prompts can be enhanced with SK's specialized knowledge. This enables a robust question-answering system to interact with users more naturally and intuitively.

    Personnel can ask questions in natural language, and GPT-4, using SK as a knowledge base, can provide accurate responses. This can be particularly useful in time-critical situations where quick access to accurate information is vital.
  3. Conflict Simulation

    GPT-4 can generate realistic conflict scenarios using real-world data integrated with SK. This can help in strategic planning and training exercises.

    GPT-4's ability to predict outcomes based on provided data can also simulate potential outcomes of different strategies. Using SK as a reference, these simulations can be grounded in the realities of the military and warfare, providing more valuable and actionable insights.

Ask Sage

"Ask Sage" is just GPT for the government. The DoD frowns upon military personnel using commercial products such as ChatGPT to accomplish work because the data is often confidential and controlled. Therefore, Ask Sage alleviates the risk of leaking data by providing GPT for classified use cases.

Ask Sage - GPT for government with CUI
Introducing Sage, the cutting-edge AI assistant designed specifically for Government teams. GPT for government with CUI / FOUO data

It also seems capable of live search and real-time API integrations, which is fantastic.

What's also interesting is comparing ChatGPT responses with Ask Sage's because Ask Sage is fined tuned with government-related data and can ingest custom data for more relevant answers for military personnel.


JADC2 Data Fabric

The US military has developed a cutting-edge command and control system called JADC2. Its purpose is to facilitate real-time sensing, analysis, and action across all domains - land, air, sea, space, and cyberspace. In the future, JADC2 will incorporate AI and machine learning to automate decision-making and provide commanders with a comprehensive view of the "battlespace."

JADC2 (credits)

Public Key Infrastructure for Defense

Public Key Infrastructure (PKI) uses asymmetric encryption, which involves a pair of public and private keys. The public key is shared freely, while the private key is kept secret. When someone wants to send an encrypted message, they use the recipient's public key to encrypt the data. Upon receiving it, the recipient uses their private key to decrypt it.

About a year ago, while pursuing a Master's degree at GWU, I had the privilege of interning at Virtru. This company specializes in data privacy, focusing on the encryption and decryption of data during transit. Founded by two brothers previously employed at the White House and the NSA, Virtru implements object-oriented encryption to protect data elements such as emails, PDFs, images, and other attachments in emails.

PKI Infrastructure (credits)

Virtru's unique key pairs ensure that when a user sends an encrypted email or file, it can only be decrypted by the intended recipient with their private key. One of the best things about Virtru is that you will always have control over your data, even after sharing it on the internet. You can easily manage and restrict access to your data, including when and where it can be viewed, like receiving an envelope you can't open until the sender grants you permission.

The administration of PKI is being integrated into JADC2. A significant issue with the JADC2 Vision involves amalgamating classified and unclassified data within a single database. Individual encryption and access controls must be implemented for shared data to unite data from various domains of different classification levels successfully.


JWCC

The JWCC is a DoD initiative that awards contracts to commercial cloud services providers like AWS, Google, Microsoft, and Oracle. These providers offer accessible cloud capabilities and services at all levels, from headquarters to the tactical edge. The JWCC ensures the military has advanced digital tools for seamless operations.

Department of Defense Announces Joint Warfighting Cloud Capability Procurement
The Defense Department announced the awarding of the Joint Warfighting Cloud Capability contracts to Amazon Web Services Inc. , Google Support Services LLC, Microsoft Corporation and Oracle.

The initiative uses an IDIQ contract vehicle to simplify the procurement of cloud services at competitive or better rates. This method enables the U.S. military to obtain a variety of capabilities through a single contract, such as:

  • Global accessibility: Ensuring that the U.S. military can operate anywhere, at any time.
  • Resilient services: Providing reliable and uninterrupted services for critical military operations.
  • Centralized management and distributed control: Offering an optimal balance between centralized oversight and distributed execution.
  • Advanced data analytics: Empowering the military with data-driven insights for strategic decision-making.
  • Fortified security: Safeguarding sensitive military data with robust security measures.

I feel fortunate to have attended the Department of Defense Azure Summit. The event highlighted numerous cutting-edge technologies and capabilities that have the potential to impact military operations significantly. As we enter this new phase of digital warfare, it's evident that the amalgamation of AI, cloud technology, and military strategy will continue to transform the defense sector in unprecedented ways.

Kyle Lyon Twitter

I'm Kyle, a Data Scientist in DevSecOps contracting with the U.S. Space Force through Silicon Mountain Technologies.

Comments


Related Posts

Members Public

On "Showing Your Work"

How can you showcase the sheer brilliance of your statistical approach in a way that's both informative and easy to follow?

On "Showing Your Work"
Members Public

Understanding Responsible ML

With great power comes great responsibility, especially when it comes to machine learning.

Understanding Responsible ML
Members Public

All Roads Lead to Probability

Probability and statistics are the basis of how we deal with the uncertainties of our world.

All Roads Lead to Probability