What are all these people doing here?��Some thoughts on AI, Humans and C2 Research

Richard Ellis

RKE Consulting

ICCRTS, Stockholm, 2025

AI and C2 – Aspirations and Assumptions

"[This] SDR [Strategic Defence Review] will transform our Army … by combining the future technology of drones and AI with the heavy metal of our tanks and artillery"

John Healey

UK Defence Secretary

"Imagine a world where combatant commanders can see everything they need to see to make strategic decisions … [and] the turnaround time for situational awareness shrinks from a day or two to 10 minutes.

Craig Martell

US DOD Chief Digital and AI Officer

"AI is perhaps the most transformative, ubiquitous and disruptive new technology with huge potential to rewrite the rules of entire industries...

UK Defence AI Strategy

"The rapid development and wide applications of artificial intelligence (AI) technology in the military domain has reshaped future combat paradigms, while presenting potential challenges to international peace and security“

Ministry of Foreign Affairs

People’s Republic of China

"Advances in data-to-decision technology, powered by artificial intelligence, machine learning and human-machine teaming offer the potential to transform sense-making and decision-making.“

UK JCCN 1/25

Developing Command and Control

Does our AI C2 research match these grand aspirations?

• AI C2 research tends to be:
• Focused on a small element of a C2 process
• Always in support of a human operator
• Assumes no change in the overall process

​

• There is a risk that we will:
• fail to explore how multiple AI nodes or agents could work directly together to produce a combined effect.
• miss the opportunity to change C2 processes in a way that would exploit the potential of new technology such as AI.
• miss opportunities to radically reduce the staffing of HQs through replacing some (or all) staff roles with AI-based capabilities

​

​

Why are we so cautious in our approach?

• General negative worldview of AI
• Long history of underestimating computers and AI
• Policy constraints
• Limited or outdated understanding of AI
• Unwarranted beliefs in the necessity for human input
• Genuine AI shortcomings

​

​

General Mistrust of AI

Policy Constraints

• Policies for the safe and ethical use of AI have emphasised the need for human control.
• 2025 “Paris Declaration on Maintaining Human Control in AI enabled Weapon Systems: “Consistent with our commitment to ensure responsible application of AI in the military domain, we will not authorize the decision of life and death to be made by an autonomous weapon system operating completely outside human control and a responsible chain of command.”
• UK Policy (eg JSP 396) emphasises the human centric approach, with AI intended to augment and empower human capabilities, rather than replace them
• In the C2 space should we be considering less stringent guardrails, in order to maximise the benefits of AI?

​

Perceived need for humans

Heard at ICCRTS 2024:

​

“AI might be clever, but there are some things [in C2] where a human will always be required.”

​

“[C2 activity] requires a type of human creativity that AI will never be able to deliver.”

​

“AI makes errors that a human would never make.”

​

Recent AI Advances

“We can confirm that Google DeepMind has reached the much-desired milestone, earning 35 out of a possible 42 points — a gold medal score. Their solutions were astonishing in many respects. IMO graders found them to be clear, precise and most of them easy to follow.”

​

IMO President Gregor Dolinar

​

Genuine Shortcomings

Humans v AI – What should we think about?

• Bias
• Correctness
• Transparency and Explainability
• Ability to Communicate
• Cost of Development
• Flexibility and Breadth
• Processing Speed and Volume
• Maintenance and Support
• Uptime (percentage availability)
• Maintaining output under stress
• Robustness
• Cost of Loss

​

A more balanced view?

• When compared on a more objective basis, the disparity between humans and AI diminishes.
• In many areas, AI is a clear winner today, and will continue to improve.

​

Escaping the Mindset��The Zero Person Headquarters

​

A Thought Experiment

A Thought Experiment – the Zero Person Headquarters

• Guiding Questions
• Could this work?
• If not, why not?
• Where would people be required?
• Today? Or always?

ZPHQ – Description and Assumptions

• Operational/ Higher Tactical HQ
• National or Multinational environment
• All HQ C2 functions carried out by AI and computers
• Standard communications with:
• Superior HQs
• Peer organisations and HQs
• Subordinate HQs
• Third parties (host nation, NGOs, press)
• Cloud deployed with communication channels as required

​

ZPHQ – Why would we want one?

• Speed of response
• Flexibility
• Ability to incorporate large datasets
• Rapid “deployablity”
• Reduced logisitics footprint
• Reduced human vulnerability

​

ZPHQ – Would it work? If not, why not?

• Challenges:
• Leadership
• Legal responsibility
• Ethical judgement
• Public communications and Trust
• Diplomacy and relationships with other authorities and organisations
• Scalability and interoperability
• Loss of tacit human knowledge
• Impact on career structures
• Support to AI infrastructure
• Need for new AI capabilities

​

ZPHQ as a thought experiment – How would we use it?

• An opportunity to:
• Think beyond current process, structure, constraints
• Explore the balance between humans and AI
• Identify genuine improvements in C2
• Identify AI requirements and constraints
• Basis for:
• Personal reflection
• Discussion workshops
• Overall feasibility
• Specialist input
• Consideration at sub-HQ level? Could this work for J2? J5? J3?
• Stepping off point for practical experiments

��AI, Humans and C2 Research

Summary

Summary

AI is needed to enable transformative changes in military C2 capability.

to focus our C2 research

​

A final thought…

What are all these people doing here?

Thank you for listening.