John R Emery (PhD Political Science @UCIrvine. Stanton Nuclear Security Fellow @stanfordCISAC) has published a new paper “Probabilities Towards Death: Bugsplat, Algorithmic Assassinations, and Ethical Due Care” in @CritMilStudies which discuss drones, ethics, collateral damage estimation algorithms, & airpower.

Original Tweet thread:



In conversation with John on Twitter, Matthew Ford explored a numbers of the paper’s ideas and themes as related to the concepts of #radicalwar

Comment #1 – Battle damage assessment methodologies

My initial thoughts – just on reading the abstract mind – were related to how this maps to the efforts put in by @airwars.


Without wanting to speak for Airwars, I understand that they successfully managed to persuade Congress that when it came to the bombing campaign in Syria, collateral damage assessments were poor.


Empirical data collection showed that targeteers could not account for civilians who (broadly speaking) were hiding in the basements of buildings.

Persistent surveillance didn’t help ‘cos people remained hidden over the long term.


So there was a mismatch between the casualty reports from in the field and those being reported by the military.

Effectively Airwars exposed this mismatch as an example of poor battle damage assessment methodology.


And forced the US military to take BDA and civilian casualties more seriously.

On the face of it, your paper seemed to me to be the flipside of the Airwars analysis.


And it stands in very stark contrast to the way that the RAF have gone about civilian casualties from their bombing campaign in the Syrian Civil War, where if you remember, they incredulously owned up to only 1 civilian casualty.


In this case it seems to me that by choosing not to look, you could deny anything bad was happening.


So, that was my initial set of thoughts – now to go read about the socio-technical construction of battle damage assessment!


Comment #2 – The role of infinite: data, potenial & targets

n=all: statistical death sentencing.

Some thoughts about it based on what I’ve just been working on for #radicalwar with @andrewhoskins

Good section. It seems to me that big data means infinite potential.

Infinite potential means infinite targets.

That implies that the way the military construct the idea of the enemy has changed. I think you discuss this is suitably fine detail.

But the key to constructing infinite targets is access to the digital archive.

If you have all the datapoints ever collected in a searchable archive then you have the capacity to continuously churn through the data to find the enemies you need to keep the whole thing on the road.

In effect then we have constructed infinite enemies.

This has worked very well for those who need to keep the Terror going – by bringing the GWOT home to American streets – so as to concentrate power.

The digital archive makes it possible to constantly find the enemy within, to find a conspiracy & the headless hydra.

It seems to me that this military strategy has now been given very real political teeth. And that this underlies much of the dysphoria that people are experiencing.

But your article also made me think about the messiness of information infrastructures in a military context. And that led me to think about Jon Lindsay’s book on Information Technology and Military Power.

IMO he’s more interested in optimising military power rather than explaining its syndromes.

& that made me wonder about the messy emergence of Skynet.

By that I mean, I wonder about the rhetorical moves that are made within military institutions to explain why something “works”

So I wonder whether you have any thoughts about the evolution of this socio-technical system and the rhetorical redescriptions it goes through as it enrolls and reframes targeting practices.

Originally tweeted by Dr Matthew Ford (@warmatters) on October 15, 2020.