DARPA will train AIs to ace close-combat aerial dogfights

AI systems will aid human fighter jet pilots in performing "split-second manoeuvring" when in the danger zone

Armed fighter jet

US military R&D arm, the Defense Advanced Research Projects Agency (DARPA), is aiming to train artificial intelligence (AI) to autonomously engage in close-range aerial duals.

At the moment, no AI system can outwit a human fighter jet pilot in high-speed, high-G dogfights, despite AI being able to trounce humans in things like professional poker, computer gaming, and in the ancient Chinese board game Go.

But DARPA's Air Combat Evolution (ACE) programme aims to change that by adding more AI-powered technology into fighter jets, which will work in collaboration with human pilots in dogfighting scenarios.

"Being able to trust autonomy is critical as we move toward a future of warfare involving manned platforms fighting alongside unmanned systems," said Lieutenant Colonel Dan Javorsek, ACE program manager at DARPA's Strategic Technology Office.

"We envision a future in which AI handles the split-second manoeuvring during within-visual-range dogfights, keeping pilots safer and more effective as they orchestrate large numbers of unmanned systems into a web of overwhelming combat effects."

By training AI in the rules of dogfighting in the same way new fighter pilots are taught, the ACE programme hopes to have systems that can take care of things such as avoiding fire from enemy jets. In doing so, pilots can apply their abilities to situations unfolding in a larger aerial battle, such as firing weapons and executing commands in the midst of combat.

The use of AI in this fashion is part of DARPA's aim to create "mosaic warfare" by which military combat is conducted by both human-controlled machines and AI-powered systems.

This approach plans to have simple smart systems networked together, rather than have the military create a single autonomous advanced platform or war machine.

The idea is that if one part of such an AI-centric network is destroyed, the rest can still work as a whole and that expendable part can be replaced and, as such, make combat more lethal for the US military's enemies while cheaper for it in terms of human and monetary costs.

"Linking together manned aircraft with significantly cheaper unmanned systems creates a "mosaic" where the individual "pieces" can easily be recomposed to create different effects or quickly replaced if destroyed, resulting in a more resilient warfighting capability," DARPA explained.

With the ACE programme, DARPA will start off training the AI systems in the basics of dogfighting under the watchful eye of instructors. Only when human pilots are confident in the AI development will such autonomous systems be put into more challenging scenarios.

"Following virtual testing, we plan to demonstrate the dogfighting algorithms on sub-scale aircraft leading ultimately to live, full-scale manned-unmanned team dogfighting with operationally representative aircraft," Javorsek added.

While the use of AI in air combat is an example of digital transformation at its most cutting edge, there are some who oppose the idea of autonomous systems being used in warfare.

A backlash from its workers saw Google pull out of its involvement in the Pentagon's Project Maven, which planned to use AI technology to improve drone performance.

Conversely, Microsoft has remained in staunch support of supplying its technology, including AI tech, to the US military, even though chief executive Satya Nadella has touted the need for an ethical approach to AI development.

Featured Resources

Security analytics for your multi-cloud deployments

IBM Security QRadar SIEM solution brief

Download now

Five reasons to move to the cloud

Join the enterprises moving their workloads to the cloud

Download now

Architecting hybrid IT and edge for digital advantage

Why business leaders should consider a hybrid IT strategy

Download now

Six reasons to accelerate remote asset monitoring with AI

How to optimise resources, increase productivity, and grow profit margins with AI

Download now

Recommended

How to become a machine learning engineer
Careers & training

How to become a machine learning engineer

23 Dec 2020
Data science fails: Building AI you can trust
Whitepaper

Data science fails: Building AI you can trust

2 Dec 2020
MLOps 101: The foundation for your AI strategy
Whitepaper

MLOps 101: The foundation for your AI strategy

2 Dec 2020
Realising the benefits of automated machine learning
Whitepaper

Realising the benefits of automated machine learning

2 Dec 2020

Most Popular

Mysterious Silver Sparrow malware hits 30,000 macOS devices
malware

Mysterious Silver Sparrow malware hits 30,000 macOS devices

22 Feb 2021
IBM reportedly mulls sale of Watson Health business
mergers and acquisitions

IBM reportedly mulls sale of Watson Health business

22 Feb 2021
Microsoft to launch standalone Office 2021 suite
Microsoft Office

Microsoft to launch standalone Office 2021 suite

19 Feb 2021