Embodied artificial intelligence and jus ad bellum necessity: influence and imminence in the digital age

• Author: Francis Grimal/Michael J. Pollard
• Position: Reader in Public International Law, University of Buckingham, UK/PhD in Public International Law, University of Buckingham, UK
• Pages: 209-275

EMBODIED ARTIFICIAL INTELLIGENCE AND

JUS AD BELLUM NECESSITY: INFLUENCE AND

IMMINENCE IN THE DIGITAL AGE

FRANCIS GRIMAL AND MICHAEL J. POLLARD*

ABSTRACT

“Prevention is better than the cure.. .” — Erasmus, 16th Century

1

See, e.g., ROYAL COLLEGE OF NURSING, https://www.rcn.org.uk/get-involved/campaign-

with-us/prevention-is-better-than-cure (last visited Jan. 10, 2022) (writing that “[t]he phrase

’prevention is better than cure’ is often attributed to the Dutch philosopher Desiderius Erasmus

in around 1500. It is now a fundamental principle of modern health care and inherent within

health and social care strategies across the UK”); see also Prevention Is Better Than Cure, THE

OXFORD DICTIONARY OF PHRASE AND FABLE, https://www.oxfordreference.com/view/10.1093/

acref/9780198609810.001.0001/acref-9780198609810-e-5664 (last visited Jan. 10, 2022).

In “re-opening” the classic debate surrounding a state’s wider right of self-

defense (in light of emerging technologies, and via the “lens” of influence com-

munications), the authors controversially “close” the following discussion in

favor of allowing Embodied Artificial Intelligence (EAI) to lawfully authorize

pre-emptive acts of self-defense in response to non-imminent threats of a grave

use of force. The authors provide a twofold justification for adopting this highly

provocative stance.

First, they argue that the introduction of EAIs will facilitate a unique recali-

bration of the necessity and last resort requirements of self-defense which would

“enable” certain pre-emptive actions to be re-categorized as “anticipatory.”

Secondly, the authors contend that because EAIs will be able to “compute” post-

bellum considerations as part of their preparatory calculations, the potential

unlawfulness of pre-emptive actions are further mitigated. In short, the utiliza-

tion of EAI’s will ensure that a greater range of humanitarian protections can

* Francis Grimal is a Reader in Public International Law, University of Buckingham, UK, and

Michael J. Pollard PhD in Public International Law, University of Buckingham, UK. The authors

would like to extend their sincerest thanks to Professor Christopher Waters, University of

Windsor, Ontario, Professor Dr. Tom Ruys, Ghent University, and Lieutenant Colonel Hamish

MacMillan of the U.K. Ministry of Defence Joint Information Activities Group for all their

considerable advice and invaluable feedback throughout the preparation of this Article. Finally,

the authors would like to extend their deepest gratitude to all at Georgetown Journal of International

Law, to Madeline Bauer Editor-in-Chief and, in particular, to Miles Malley and Sophie Mehta for

all their considerable suggestions and recommendations during the edits—the authors are

extremely grateful & are delighted to be working with the Journal on this third installment of the

series. Please note however, the views expressed herein are entirely those of the authors. V

C 2022,

Francis Grimal & Michael J. Pollard.

1.

209

be provided to the civilian population when future acts of self-defense are

deemed necessary.

I. INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210

II. THE JUS AD BELLUM AND JUS IN BELLO LEGAL FRAMEWORK . . . . . 220

A. Prohibition on the Use and Threat of Force . . . . . . . . . . . . . 222

B. Self-Defense under International Law . . . . . . . . . . . . . . . . 224

C. Jus in Bello Self-Defense . . . . . . . . . . . . . . . . . . . . . . . . . . 229

III. INFLUENCE COMMUNICATIONS, PROPAGANDA AND SELF-DEFENSE . . 234

A. Overview of Influence Communications. . . . . . . . . . . . . . . 236

B. Influence Communications and International Law. . . . . . . 239

IV. TEST AND SCENARIOS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247

A. The Authors’ Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248

B. Scenario . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252

C. Analysis 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254

D. Analysis 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257

E. Analysis 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260

F. Analysis 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263

G. The Proliferation of EAI Technology. . . . . . . . . . . . . . . . . . 264

V. WIDER IMPLICATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265

A. Recalibration of Collective Security. . . . . . . . . . . . . . . . . . . 266

B. Recalibration of Extra Charter Exceptions (Humanitarian

Intervention and Responsibility to Protect) . . . . . . . . . . . . . 269

VI. CONCLUSION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271

I. INTRODUCTION

In a briefing on July 28, 2021, a top U.S. military commander

explained how the Pentagon is utilizing Artificial Intelligence (AI) to

achieve information dominance and enable military planners to look

far beyond the temporal urgency of seconds, minutes or hours, to

instead predict a behavior or action in the coming days ahead.

2

See Gen. Glen D. VanHerck, NORTHCOM Commander, NORTHCOM Commander Glen

D. VanHerck Conducts Press Briefing on North American Aerospace Defense Command and

U.S. Northern Command Global Information Dominance Experiments, (July 28, 2021)

(transcript available at https://www.defense.gov/Newsroom/Transcripts/Transcript/Article/

2711594/northcom-commander-gen-glen-d-vanherck-conducts-press-briefing-on-north-america/)

[hereinafter GIDE].

As a

result, in the following analysis (the third in a series of articles),

3

the

2.

3. For previous analyses see Francis Grimal & Michael J. Pollard, “Embodied AI” and the Direct

Participation in Hostilities: A Legal Analysis, 51 GEO. J. INT’L L. 513 (2020) [hereinafter Grimal &

Pollard (2020); Francis Grimal & Michael J. Pollard, The Duty To Take Precautions in Hostilities, and

GEORGETOWN JOURNAL OF INTERNATIONAL LAW

210 [Vol. 53

authors use the exclusive nature of Embodied AI (EAI) to revisit the

concept of imminence within the jus ad bellum, and recalibrate the

“necessity threshold” of self-defense. The authors’ proposed recalibra-

tion would lawfully permit a state to act pre-emptively against a non-

imminent and latent threat of a grave use of force.

To reach such an undeniably controversial conclusion,

4

the authors

begin by identifying the previously anonymous concept of “perpetual

self-defense” — a notion which is squarely routed within the ever-

expanding domain of influence communications and propaganda.

5

As

its name suggests, “perpetual self-defense” is the recognition that all

states are inherently (and as a default setting) “on the defensive” (be it

DEFCON 4 or DEFCON 1)

6

particularly, in the context of influence

communications and propaganda where states are constantly attempt-

ing to manipulate and even coerce their adversaries in the short, me-

dium, and long term. Undeniably, the reader of this Article may

immediately raise “objection” and quite rightly assert that while the

recourse to influence communications and propaganda is an unsightly

form of statecraft, it is one that falls outside the stringent regulation of

the ad bellum framework.

The authors counter this, however, by identifying a second previously

anonymous concept — that of “self-defense by proxy.” “Self-defense by

proxy” is used by the authors to denote a form of indirect pre-emptive

self-defense that is best achieved by utilizing influence communica-

tions. Perhaps somewhat analogous to an indirect use of force, self-

defense by proxy is where the authors envisage the use of influence

communications and propaganda as part of a long-term strategy to

destabilize a state adversary.

By way of example, one might envisage a sliding scale where, at the

minor end, a state could choose to “bombard” its adversary, State B,

with “anti-vax” propaganda in relation to the uptake of immunizations

against Covid 19 (though undeniably harmful, such an act is one which

clearly does fall outside of the ad bellum remit). In contrast, however, at

the Disobeying of Orders: Should Robots Refuse?, 44 FORD. INT. L. J. 671 (2021) (considering EAIs in

greater detail in Part II) [hereinafter Grimal & Pollard (2021)].

4. Regarding the controversial nature of this conclusion, see the authors’ discussion infra

Part II.

5. Noting this concept is considered in greater detail throughout and specifically discussed

infra Part IV.

6. DEFCON is the United States’ Nuclear Defense Condition System. There are five levels of

military ‘readiness’; 5, Low/ Normal; 4, Above normal; 3, Airforce ready to mobilize in 15

minutes; 2, Armed forces primed and ready to be deployed within hours; 1, Maximum readiness,

capable of acting/ responding immediately. For a useful discussion/ analysis see, e.g., Scott D.

Sagan, Nuclear Alerts and Crisis Management, 9 INT’L SEC. 99, 100-02 (1985).

INFLUENCE AND IMMINENCE IN THE DIGITAL AGE

2022] 211