Artificial Intelligence in Space: Providing Near Real-Time Threat Analysis

Approximate Reading Time: 12 Minutes
By Justin M. Jettòn


In 2019, the global space economy was valued at over 423 billion dollars (USD) and will continue to trend upward for the foreseeable future. Within the space economy, the employment and use of satellite systems constitute 271 billion dollars (USD), or 65% of the total space economy. This economic growth appeals to government and commercial entities, which have shifted investments from other areas into space. This shift, coupled with the reduced cost of entry into space, has led to an explosion of active satellite systems. Operational satellites have doubled over the last two decades, with an expected additional increase of 3,000 busses by 2026. As more satellite systems come online, the global reliance on these systems will also increase. This reliance is due to commercial and government organizations shifting services to space. These services include critical infrastructure, national security, economic, and societal support systems. As these services move towards space, they create vulnerabilities that, if exploited, can have a significant impact on the economic health and security of the United States.

These newly created vulnerabilities have not gone unnoticed by enterprising countries looking to become the dominant force in space. China is one example of many nation-states looking to counter the United States military capacity in space through various offensive capabilities, according to the China  Security Report 2021. China is not alone in its intent to meet or exceed US capabilities in space. Adversaries and allies alike are looking to leverage the space domain for economic gain and national security, creating the competitive and contested environment in which the United States now finds itself. The United States no longer has the luxury of operating in an open, uncontested space environment. For the United States to remain competitive and ensure freedom of movement, it must utilize and employ advanced technologies such as artificial intelligence (AI) to its advantage. Known or assessed variables, such as counter-space weapons location, energy weapons range and power, and environmental factors, can be utilized to feed algorithms to provide accurate assessments. AI can be used as a near real-time threat evaluation model to help drive the decision-making process at the speed of relevance. One such viable AI model is the Markov Chain. To best understand the need for this model, it is pertinent also to understand if there is a conceivable need for AI in space defense. This article will explore the possible counter-space threats, if nation-states have top talents in the field for research and development of counter-space systems, and what a Markov Chain could help provide in space defense. 

Counter-Space Threats

A counter-space threat is any entity that can threaten a space system. This includes any part of a space system from the ground station, information networks, or satellites. Counter-space threats are nothing new, but they have seen an increased resurgence over the last decade as governments seek to utilize space and deny others from using the domain. Weeden and Samson break down offensive counter-space into five categories; direct-ascent, co-orbital, electronic warfare, directed energy, and cyber. Out of the five categories listed, directed energy laser weapons are the optimal use case for a Markov Chain model. This is due, in part, to the limited variables associated with laser weapons compared to other threats such as electronic warfare or cyber. In addition to the focus on laser weapons, this article will highlight the capacity of nation-states to pull top talent within their own country and the threat posed by directed energy weapons (laser).

Directed Energy Weapon – Laser

To understand how AI can provide automated space defense, you must understand how a counter-space directed energy laser works and its associated variables. The variables would ultimately feed into an automated time threat analysis model, and without them, the algorithm would not function as designed. When electrons in glass, crystals, or gases absorb energy from an electrical current and become excited, they create a laser. The power put into exciting the electrons determines a laser’s strength. Thus, a lower-powered laser held in your hand will not be as powerful as a laser connected to a significant energy source. The more powerful the laser, the further it can travel and the more damage it imparts on a target. Power capacity is only one of the factors that can affect the capability of a laser system. Another factor is the weather.

Clouds and precipitation can also significantly affect the propagation of a laser. As a laser passes through air and space, the precipitation in the atmosphere can cause a laser’s energy to be absorbed by water and air molecules, decreasing the laser’s energy output as it travels. This propagation can severely impact the strength and distance a laser will travel. Though brief, this understanding of laser systems and some factors that can limit their effectiveness can feed into an automated probabilistic algorithm to help assess the potential threats lasers have against satellite systems. With a baseline understanding of laser capabilities and limitations, it is also essential to explore nation-states likely to develop laser-based directed energy weapons.

Talent in Demand

An indicator of a country’s technical mastery is through a process known as bibliometrics research. For this article, bibliometric research provides an essential tool for identifying nation-states with the likely capacity to develop directed energy counter-space weapons. Bibliometrics research analyzes publication volume, journal impact factors, most cited articles, preferred methods, and represented countries. Within the umbrella of bibliometrics is a method known as the 3-F method. The 3-F method is used for this article to provide insight into top talent indicators from nation-states. The 3-F method culminates in the brain gain index, determining if a country has enough top talent within its geographical boundaries.

OTH, Emerging Security Enviroment, Multi-Domain Operations, Joint All-Domain Operations

Linjia and his team, who study technology innovation and talent management at China’s National Academy of Innovation Strategy, developed this specific formula for calculating top talent in a nation-state. The equation breaks down into the following:  Iik is the brain gain index value of country (i) in the field (k), Twk means the number of the world’s top talents in the area (k), Tik implies the number of country’s (i) top talents in the field (k), Pw means the world population, Pi means the country’s (i) population. If Iik was more than 1, that means the country (i) has fewer top talents in the field. The 3-F method captures which countries have enough top talent in space and directed energy. Using a keyword search and frequency over a 12-month timeframe (Table 1), an analyst can identify which countries have more robust brain gain indexes and thus more substantial talent in the field (Table 2).

OTH, Emerging Security Enviroment, Multi-Domain Operations, Joint All-Domain Operations
Table 1: High Impact Keywords and Frequencies

12 of the 14 top countries have a brain gain index of less than 1, indicating that they have enough top talent in the field to both develop and deploy directed energy counter-space capabilities internally. The countries with a brain gain index greater than one will most likely bring in foreign talent or purchase equipment from other countries to develop counter-space directed energy lasers. This method highlights those nation-states outside the United States that can likely develop and train top talent in counter-space weapons, further highlighting the need to focus on advanced technologies to secure the space domain.

OTH, Emerging Security Enviroment, Multi-Domain Operations, Joint All-Domain Operations
Table 2:  Brain Gain Index of Top 14 Countries

As mentioned earlier, the 3-F model utilizing the brain gain index is a viable method to explore likely nation-states with top talent. Still, it is not the only model the US should use for insight into nation-states’ capabilities. The Defense Intelligence Agency (DIA) released a report that highlighted both Russia and China were actively developing directed energy counter-space threat systems. As shown in Table 2, according to the 3-F model, China does not have enough talent within its borders to build directed energy counter-space weapons. However robust, the 3-F model cannot address all variables and looks only at published articles and research. China and other nation-states with higher brain gain indexes could have prevented journal publications or could import talent from other nation-states for research and development.

Markov Chain Model

The use of the Markov Chain probabilistic model for automated threat analysis has shown success in recent years and may prove viable in analyzing laser-based counter-space weapons threats. A study conducted in South Korea on the viability of using Markov Chains to predict and evaluate enemy air fighters showed a higher rate of accuracy when compared to previous AI models, highlighting its robustness in a near-real time environment. The Markov Chain model is used to determine the probability of a future state based solely on the present state, not on a sequence of consolidated events from the past. This lack of reliance on historical data makes a Markov Chain model effective without relying significantly on ample data storage and computational power. The lower computational power and storage requirements and its proven role in cyber defense assessment make the Markov Chain an ideal model for space-based threat assessment.

OTH, Emerging Security Enviroment, Multi-Domain Operations, Joint All-Domain Operations

P is the probability where Xn+1 identifies the future state while x is the future state itself. Xn is the current state, while xn is the state itself. Using Figure 1 as an example, the probability equation would look like P(X4 = A|X3 = B)=0.6 stating that based on the observed attributes,  there is a 60% chance that the state will transition from state B to state A.

OTH, Emerging Security Enviroment, Multi-Domain Operations, Joint All-Domain Operations
Figure 1:  Visual Representation of 3 state Markov Chain

The Markov Chain model for space threat assessment would have four states associated with it; threatened, not threatened, threat increase, and threat decrease. As a space asset moves through the threat ring of a laser system, the attributes shown in Figure 2 will drive the Markov Chain prediction and provide a threat assessment based on those observed attributes. Power of the laser system and weather are two features that would feed the algorithm. Power would help create an initial threat distance, while weather conditions would drive the real-time threat distance of the laser system. As the weather changes in real-time, the threat posed by laser systems would also change. Since the Markov Chain is iterative, it would continue to assess the probability of a threat as observed attributes change. This provides a dynamic near real-time threat assessment for space assets.

For a simple example, using the threatened (D) and threat decrease (A) states, it becomes clearer how a Markov Chain would function. A satellite system has moved overhead of a directed energy threat with the previous Markov Chain iteration moving from increased threat to threatened. As the satellite continues its orbit, it increases its distance from the weapons system. As the Markov Chain continues its iterative assessment, it would provide a 90% probability of decreased threat as it continues to distance itself:

OTH, Emerging Security Enviroment, Multi-Domain Operations, Joint All-Domain Operations
OTH, Emerging Security Enviroment, Multi-Domain Operations, Joint All-Domain Operations
Figure 2: Overview of Process

Way Ahead

This paper only provides a brief overview of what an AI model for space threat assessment might look like. Some of the areas not addressed include ingesting the various attributes required to drive the threat assessment, how this model would scale beyond laser weapons, and accurate weather impacts to laser systems. These three areas are vital for the success of this model or any AI model expected to help provide threat assessments to space systems. On top of looking further into the areas addressed above, current initiatives in AI must incorporate proper data management, policies, engineering and ensure that AI systems can start small but expand and grow as the needs arise.


Space is complex and becoming increasingly crowded. Commercial and government entities are vying for a piece of the space economy, shifting the United States’ position. Numerous countries have the talent and knowledge to operate in space and develop weapons to counter the United States’ operation and interest in the domain. This has shifted space into a contested environment. The US must look at space differently, adapt, and develop tools to ensure that the US, its allies, and their economic interests continue to have freedom of maneuverability. To do this, the US must adjust to a dynamic and ever-changing environment at speeds not operated at before. One method to meet this need is developing AI tools that enable a faster decision-making process and allow the US to perform at the speed of relevance.

Justin Jettòn is an intelligence officer in the United States Air Force. He recently graduated from the Junior Officer Cryptologic Career Program (JOCCP), where he focused on data science applications in cyber defense and cyber vulnerabilities to space systems. He is currently working on his Ph.D. in Information Technology.

Contact Information:

The views expressed are those of the author and do not necessarily reflect the official policy or position of the Department of the Air Force or the US Government.

Featured Image Source:

OTH, Emerging Security Environment, Multi-Domain Operations
Print Friendly, PDF & Email

Leave a Reply