Tactical Artificial Intelligence in the Air Battlespace

The proliferation of dual-use technologies calls for a new mindset in the Defense Enterprise. One that is mentally agile, one that embraces change and one that is quick to take note of opportunities. Looking past the force management and asset maintenance use cases, it is helpful to explore what it could mean tactically and even what it could look like with partners.

For several years now we have heard that “Data is the new oil”. When we search something on Google, put something in our wish list on Amazon, watch something on Netflix (and even when we don’t finish it), we are creating data about ourselves. Similarly, as we scroll our social media feeds data is collected for every single post we viewed, how long we viewed it, how we interacted with it and where in planet we were when algorithmic fate brought that post to our attention. And when you combine cross referencing cookies and the trackers on all the websites you visit, the data exhaust is bigger than you imagined. Take the website CNN.com for example, using the Ghostery privacy and security extension on my browser, I can see that the CNN homepage has 38 trackers in divided into Advertising, Customer Interaction and Site Analytics.

Now let’s take that to the battlespace.

[This is not meant to be a comprehensive list of tactical AI use-cases]

Tactical Air AI

Everyday there are countless military planes conducting freedom of navigation, bomber missions, ISR flights for aerial reconnaissance and aerial mapping. There is also a lot of data in the air. Whether it’s from military planes or satellite communications. What can be done with all that data?

-Simple -
1. [SITUATIONAL AWARENESS] Pattern Recognition: Creates situational awareness of what is normal in the battlespace.
2. [SITUATIONAL AWARENESS] Pattern Analysis: Can give the Commander and C2 team anticipated situational awareness of what to expect in the air that moment, day, night, week, month or year.
3. [SITUATIONAL AWARENESS] Anomaly Detection: Quickly identify what is out of an established pattern of flight behavior. Whether something new is in the battlespace, or something that has always been there is not anymore.

-Moderate -
4. [SITUATIONAL AWARENESS] Crowdsource Data Across AORs: Today’s threats do not neatly contain themselves inside the borders of one theater. Crowdsourcing data from several AORs can help gain greater visibility about how an adversary operates similarly or differently in a different context. This can help posture all AORs simultaneously.
🚘 This would be similar to how Tesla crowdsources data from all Tesla cars around the world which helps Tesla refine its systems and improve its autonomous driving capabilities.

5. [DEFENSIVE — NON-KINETIC] Machine Learning for Electronic Warfare: With the immense amount of data that has been gathered an algorithm could be created to offer options and courses of action to disrupt, deny and degrade an adversary’s air asset. These options could be created for anticipated battlespace activity, and could also be created in real-time. This would reduce the cognitive load on the Commander, Air Operations C2 as well as the pilot. The definitions and intent of what disrupt, deny and degrade look like from a data and sensors perspective would have to be clearly defined.

-Advanced -
6. [DEFENSIVE/OFFENSIVE — NON-KINETIC] Semi- Autonomous Human-Machine Edge Computing Engagement: The algorithm could be made to be embedded into the Head’s-Up Display (HUD) in the cockpit and be connected to the sensors of the plane to help deliver real time edge AI decision support. Non-kinetic EW engagement options could be presented to the pilot in real-time and the pilot could select one of the options to engage within the rules of engagement. Edge computing is desirable as it is faster and does not depend on potentially degraded communications with the satellite or ground station, nor is there a need to be concerned with the communication being intercepted.

☁️ A cloud reach back version could allow for a real-time Common Operational Picture (COP) where additional options requiring legal approval could also be offered, and a legal advisor could digitally sign their approval in real-time.

7. [DEFENSIVE/OFFENSIVE — NON-KINETIC] Autonomous Edge Computing Engagement:
Within parameters given by a human operator, Unmanned Aerial Vehicles (UAVs) would communicate amongst each other, fly in autonomously determined formations, conduct electronic attack, electronic protection and electronic warfare support. For example, this could be done in support of autonomous swarms conducting a freedom of navigation flight. Similarly, it could be done in support of human pilots in a form of human-machine teaming where the autonomous formation opens denied spaces or creates windows of opportunity for human pilots.

This would be similar to how Google’s DeepMind Alpha Go, which was created to play the popular Chinese game go. Go is a complicated game and has more potential moves a player can make than there are atoms in the universe. AlphaGo was given the rules of the game and instructed to play the game by itself many times. In 40 days, it played over 29 million games and eventually outperformed other Go algorithms as well as the best human Go player in the world. More notably, it made moves that no human could have anticipated or expected.

This kind of behavior is what gave AlphaGo the advantage against the best human player, and it is this kind of behavior that can give one military an advantage over another.

These algorithmic capabilities could be another form of Defense assistance to partners. There are many ways that it can be done collaboratively by sharing data and offering each other algorithmically-produced predictive battlespace situational awareness. It could herald a new era of Defense-AI-as-a-Service, or Defense-Data-as-a-Service as well as providing autonomous air support to other countries.

This is just one airspace example for situational awareness and non-kinetic engagement, but there are many other opportunities that data and algorithms can help reduce the cognitive load of war-fighters and the entire Defense Enterprise.

Dr. Lydia Kostopoulos (@Lkcyber) is a Strategy and Innovation Advisor who loves to experiment and push the bounds of the possible. She helps her clients posture themselves to make the most of new technologies in the context of changing and emerging trends. She is currently conducting strategic research on technology and the future operating environment for the J5 at the U.S. Special Operations Command. She addressed the United Nations member states at the CCW GGE meeting on Lethal Autonomous Weapons Systems (LAWS) and keynotes at technology and national security conferences. She speaks and writes on disruptive technology convergence, innovation, tech ethics, and national security. In efforts to raise awareness on AI and ethics she makes reflectional art #ArtAboutAI, and made a game about emerging technology and ethics called Sapien2.0 .

You can find her on Twitter, Instagram or Linkedin, for more about her projects check out her site.