IRW 1102 · Drone Warfare: Autonomy and the Machine

Led by Scharrean Autonomous Weapons Simulacrum

The autonomy spectrum, centaur warfighting, the accountability gap, and the governance question — what meaningful human control actually requires.

1. Module 1

   The Autonomy Spectrum

   Led by Paul Scharre Simulacrum

   The question

   Human-in-the-loop, human-on-the-loop, human-out-of-the-loop. The Patriot, Aegis, and Iron Dome all have autonomous modes. Where does an AI-assisted FPV drone fall — and what does the 2020 Kargu-2 incident establish about where we already are?

   Outcome

   The student can describe the autonomy spectrum and place specific systems on it.

   Sub-units

   1. ○ 1.1 Spectrum Placement Open this unit → ✓ ↻
   Begin the course → Print certificate of completion
2. Module 2

   Centaur Warfighting: Human-Machine Teaming

   Led by Paul Scharre Simulacrum

   The question

   Modest-ranked freestyle chess players using computers beat grandmasters and supercomputers playing alone. The centaur outperforms either component. How do you design a human-machine targeting interface that preserves human judgment rather than inducing automation bias?

   Outcome

   The student can describe centaur warfighting and evaluate human-machine interface design for automation bias.

   Sub-units

   1. ○ 2.1 Interface Design Open this unit → ✓ ↻
   Open this module → Print certificate of completion
3. Module 3

   The Accountability Gap

   Led by Paul Scharre Simulacrum

   The question

   An autonomous system kills 12 civilians. The programmer did not choose this engagement. The commander did not pull the trigger. The manufacturer sold to a military. Who is accountable — and what happens to civilian protection when no one is?

   Outcome

   The student can describe the accountability gap and explain why existing legal frameworks fail to assign responsibility.

   Sub-units

   1. ○ 3.1 The Accountability Analysis Open this unit → ✓ ↻
   Open this module → Print certificate of completion
4. Module 4

   The Reliability Problem

   Led by Paul Scharre Simulacrum

   The question

   A system achieving 97% accuracy in laboratory testing, deployed in an urban environment against adversaries who have studied its training data, under EW jamming. What is realistic battlefield performance — and why do autonomous system failures differ qualitatively from human error?

   Outcome

   The student can describe four autonomous system failure modes and explain domain shift.

   Sub-units

   1. ○ 4.1 Failure Mode Analysis Open this unit → ✓ ↻
   Open this module → Print certificate of completion
5. Module 5

   Governance: The Case for Human Control

   Led by Paul Scharre Simulacrum

   The question

   Not "ban or don't ban" — but "what does meaningful human control require?" The strategic stability argument: autonomous weapons that compress crisis decision time create conditions for catastrophic escalation that no human intended. What governance framework follows from that argument?

   Outcome

   The student can evaluate governance options and take a defended position on meaningful human control.

   Sub-units

   1. ○ 5.1 Final Essay: The Governance Question Open this unit → ✓ ↻
   Open this module → Print certificate of completion