Systems Engineering in Aerospace & Defense Systems Critical to National Security, in this Age of Large Language Models (LLM), ChatGPT, and Copilot

Weapon systems critical to national security are increasingly becoming more complex and expensive. The cost overruns alone are estimated to be about half-a-trillion dollars, making it the second largest defense budget in the world. In the USA, several weapons system programs breached Nunn-McCurdy thresholds and one of them is termed the “Most troubled program in the entire DOD”. Factors that led to the critical Nunn-McCurdy breach include “Inadequate systems engineering at program inception, Block-0 software with high defect rates and Block-1 designs requiring significant rework”

“Inadequate systems Engineering” is almost always the critical factor in troubled programs. Systems Engineering, since the days of Bell Labs where the word was first coined morphed several times to the current state where it has become tool centric and lost much of its original purpose of guiding the development of systems (in particular large and complex systems like 5ESS and Aegis BMD both are 100 million lines of code class). If it does not adapt to the changing technologies, it will soon be relegated to history.

By concentrating almost exclusively on the tools like Cameo and UML/SysML artifacts which more often than not are dead ends and not seamlessly integrated into the Systems Development Life Cycle. Increasingly Systems Engineering is being lumped with Testing (Systems Engineering, Integration and Testing (SEIT)). It is interesting to note that almost always System Engineering, in particular MBSE (Model based Systems Engineering) touts as a great improvement on the document-based processes. However, almost always, the idea for a new system or capability starts with a Capability Development Document (CDD). I am yet to see where the requirements form the user starts with SysML artifacts.

Software development a.k.a computer programming also went through its own tool centric mania. However, with the advent of A.I, LLMs, ChatGPT and Copilot [28-36] the software development is witnessing huge gains in productivity and quality. Programmers are not spending their time and energy in creating UML diagrams. The LLMs can directly take the requirements from Unstructured, Uncertain, Incomplete, Imprecise, and even Contradictory (UUIIC) information from all sources (documents, audio, video, web, emails, hidden web, etc.) and generate the final systems probably using 3-D printing. This will make the current processes obsolete, including Systems Engineering and computer programming. Even though this is not yet perfect and ready for prime time, they are progressing in the right direction.

In this changing landscape, unless Systems Engineering jettisons its marbles (Cameo, SysML, Models, etc.) and try to become an integral part of the system development life cycle, soon Systems Engineers will be extinct like the office typists and stenographers of yester years. By concentrating almost exclusively on MBSE to the detriment of Modeling, Simulation and Analysis (M&SA) which is a valuable tool for the designing the system and conducting the Analysis Of Alternatives (AOA), Systems Engineering will lose its relevance.

This lecture presents the state of systems engineering in particular for the large, mission critical weapons systems, identifies the areas susceptible for improvement (adapting to the new groundbreaking and disruptive technologies) so that it will still be relevant and useful in the System Development Life Cycle. This paper will lay out the subtle course corrections that can be made by Systems Engineering to be relevant and useful.

Note: This paper does not discuss any specific details of any weapons system or program other than those published in the Government Accountability office (GAO) and Congressional Research Service (CRS) reports.