Wednesday, July 8, 2015

Ethical Weapons and Warfare in an Era of Ubiquitous Surveillance


My liberal friends sometimes ask me why I'm a member of the science fiction think tank SIGMA, under the assumption that it's a right-wing organization. (It isn't -- you should hear the arguments the group gets into over dinner, coming from all parts of the political spectrum at once.) I tell them my reasons are twofold:

First, I despise terrorists more than I distrust the American government.

Second, I want to do my small best toward making the world a better place. Most of SIGMA's sponsors are looking for new things to be afraid of. My agenda is to offer up possibilities that will result in less oppression and greater freedom.

Buried deep within the following presentation, which I made at a conference on the future of small arms, is exactly that agenda.

Ethical Weapons and Warfare in an Era of Ubiquitous Surveillance

By Michael Swanwick

Ubiquitous Computing
You probably already know about ubiquitous computing.  But just in case…  UC envisions a future in which information processing has been thoroughly integrated into everyday life.  Small, robust, and inexpensive processors will be everywhere – in cars, refrigerators, lamp posts – and networked so that as much processing power as anybody needs will be available, essentially free.  Some go so far as to imagine that fogging machines will periodically go down all major streets, spraying nanocomputers into nooks and crannies as a public service.  There are many people working right now to make this happen.

Ubiquitous Surveillance

Similarly, ubiquitous surveillance – the presence of networked cameras and recorders in every imaginable public space – is looking more and more like the wave of the present.  Britain already has a national system in place, but even in relatively libertarian America, the quantity of surveillance is mushrooming as cameras grow smaller, cheaper, and easier to install.  It doesn’t matter whether people are working on it – it’s self-assembling right now.

Ubiquitous surveillance is usually imagined in the form of a unified system rigidly controlled by a central government authority – as in George Orwell’s 1984 – and for this reason, many governments think of it as an attractive anti-terrorism tool.  In fact, it will grow from what already exists – private security cameras, traffic cameras, web cams, and the like, in layer after layer of not-always-compatible technology.
This means that the surveillance web can never be perfectly controlled.  It is a resource available to everyone.

Transparent Combat

In a future war where both sides share common territory and ubiquitous public surveillance, a uniformed conventional army must always be at a disadvantage.  Not only can the entire surveillance web not be controlled but, worse, you can never know which parts of it are available to the enemy.

Therefore you must always assume yourself to be under surveillance and act accordingly.
In such a scenario, likelihood of victory will be defined not by superiority of position or of information, but by control of the moral narrative.  The force that is popularly seen to deserve to win – either because they are more virtuous or because they are unstoppable– will be able to rely on the host population’s at least passive cooperation.

Enemy Response
One of the enemy’s primary objectives will be to get video of American soldiers committing atrocities.  If this does not happen in the normal course of war, an obvious tactic would be to provoke or trick the soldiers into war crimes.  If you go into the slums and give every kid who wants one a bright scarf in a color identified with the enemy and then offer twenty bucks to the first child who can reach a specific army checkpoint, the ensuing race will look like an assault.  If one of the slower children is carrying a disruptor device that will shut down local military surveillance cameras, you’ve got a frightening wave of blindness coming toward edgy men with weapons.  If the child with the disruptor is slow enough, the military’s own cameras can be used to capture the incident.

Imaginative Cowards
Many such ploys will be invented by self-appointed tacticians safely isolated by distance from their consequences.  Terrorism is, as a rule, an unimaginative act simply because imaginative people can visualize their own suffering and death too vividly to make good suicide soldiers.  But transparent warfare gives imaginative cowards the opportunity to actively contribute to their cause by monitoring the surveillance web for images that make American forces look bad, by searching out weaknesses in the surveillance and communication webs, and by inventing tactics, evaluating their effectiveness, and distributing this information via websites whose operators are connected only by ideology.
Thus, the advantage of conferences such as this one is negated, and the enemy can be expected to display greater ingenuity than has previously been the case.

A Camera for Every Gun (Passive Response)

In order to counter these disadvantages, the first thing to do is to place cameras on every weapon and vehicle, set to record whenever it is in use.   Initially, this will be as unpopular as the now-common practice of equipping squad cars with cameras was with the police.  In practice
however, these cameras have proved useful tools not only in preventing abuse but also in showing the police officers’ true actions in situations that otherwise looked incriminating for them.  The soldier’s gun will quickly come to be seen as a fair witness, one that accurately presents his own point of view.

Two Worlds, Three Audiences
In such a scenario, taking down parts or all of the surveillance and communications webs will be a popular tactic of the enemy, either through spoofing, hacking, or physical sabotage. Thus, the effective soldier will be one who can easily toggle between two modes of perception and command.  The first is a fully integrated hierarchic system, in which civilians with no military experience will be able to micro-manage field operations via handheld from a bar in Georgetown.  The second, when such systems are down or unreliable, is in essence pre-electronic warfare, and requires soldiers who can read a map, keep their rifles clean, and continue their mission on their own.
To complicate matters, when the surveillance web is functional there will always be three audiences for all actions:  The enemy, the host population, and civilians back home.  Each of which will use the same information very differently.           

Ethical Weapons

The soldier fighting in two worlds and with three audiences will need weapons that are both independent and ethical.

By independent I mean not networked.  This prevents a weapon from being hacked and turned on its operator, of course.  But, more importantly, the weapon will behave identically in either mode of combat, thus providing a behavioral anchor for the soldier caught in a potentially confusing situation.

An ethical weapon is one which is designed to decrease the chances of an incident which would shock the conscience of noncombatants.  Such weapons include but are not limited to those which give the soldier the option of non-lethal force; those which can assess the conditions of combat and vary the velocity of projectiles, to limit the deadliness of stray fire; those which can deny their services to unauthorized users; those which can be programmed to recognize and refuse to fire on defined groups such as women or children; and those which will not fire at friendly forces.

It is easy to imagine recognition patterns or devices that would protect all noncombatants or friendly forces from misdirected fire, but difficult to imagine any that could not be cracked and replicated by a technologically-savvy enemy, leaving soldiers with guns that will not fire at anybody.  But with good biometric programming, it should be possible to at least identify babies, infants, and children up to a given age, and protect them from direct gunfire. 

Rethinking the Rifle
For simplicity’s sake, I’m going to be talking about the rifle from this point on.  But my observations are not necessarily limited to one weapon, and it is not necessary that all these functions be combined in a single device.
Nevertheless.  If a soldier’s best friend is his rifle, shouldn’t his rifle be more like a dog?  A dog is not only fierce in combat, but loyal and a good companion as well.
A rifle’s loyalty comes in three layers.

The first is simple operation.  As rifles grow smarter, they should be capable of recognizing at a minimum their user, members of his platoon, superior officers, and other individuals they have been “introduced” to.  Depending on how it is programmed, a rifle can either refuse to fire upon them under any conditions or else warn its  user about the target’s identity upon being aimed.   It  should be capable of seizing up if handled by a stranger without permission from somebody whose authority it recognizes.
The second is security.  In a mature electronic era – one in which everybody involved has a full grasp of the technology – networked robotic devices become enormous potential liabilities not only because they can malfunction but because they leave open the possibility of being hacked by the enemy.  For this reason, it is important that a combat soldier’s rifle be neither networked nor autonomous, but isolated, alone, and reliant upon its user.
The third is psychological.  For the soldier to operate at peak efficiency, the gun must be perceived as his, an ally, rather than as a means his superiors have of maintaining control over him.  His rifle’s recordings should never be accessed lightly, but only under clearly-defined conditions.  Its programming should be as little restrictive on his actions as possible.  Nor, save in extreme situations, should the rifle overrule him.  It should always be clear that it is the soldier and not his weapon who is in command and making the decisions.

The Rifle as Companion

Long before AI becomes a reality, a rifle should be able to converse with a soldier in a simple, natural manner, offer practical advice – such as “Don’t offer to shake with your left hand” or “That water will give you dysentery” – provide rudimentary translation, and make a good guess as to the emotional state of someone being interrogated.

It  should also be able to read a soldier’s emotional state and provide appropriate counseling in non combat situations.

By this, I do not mean psychotherapy. In ancient Rome, when a victorious general rode in a triumph, a slave stood in the chariot behind him to murmur from time to time, “Remember, you are only a man.”  This worked because the general knew it to be true and understood why it was being said.

I do not presume to know what truths should be loaded into a rifle’s knowledge bank, other than that they should be capable of being edited or deleted by the soldier, and that they should be recognizably not propaganda.  But a device which knows when to crack a joke or offer commiseration (and when to be silent) will be perceived as not only a useful tool but a good companion.

The Disadvantages of Robots
I want to make the distinction between this sort of device and a robot particularly clear, because in transparent combat, robots have several drawbacks.  The first is that most people feel a primal fear of them.  Video of self-propelled machine guns moving through a village immediately burdens you with the unstoppable alien invaders narrative.  Whether you want it or not, you’ve assumed the role of Darth Vader.

The second is if only a single weapon is hacked and turned on your own soldiers, they will all henceforth distrust their own weapons.
The third is that people who would not try to kill a human being will feel no such compunction toward a robot.  A farmer will feel perfectly justified shoving a hoe into the works of an autonomous mobile howitzer that’s tearing up his fields.  A mother who thinks an urban patrol unit is threatening her child will empty the bedpan into its electronics.  Nobody will have to tell underage boys to throw rocks at any robots they see.  If the robots fire upon any of them, the video will be posted on YouTube within the hour.  If they refrain but are destroyed, that video will be made available wherever bored young men are looking for something fun to do.

Man and Weapon as Cyborg
This is why the rifle should be designed as a sidekick, not as a player.  Its autonomy should be limited and its intelligence should be primarily advisory.  The gun should be completely dependent upon the man.
Simply by being a member of the military a soldier is always networked, even when he is alone and unable to communicate.  His weapon, therefore, need only answer to him. 
When the soldier is well trained and his weapon supports him by protecting not only his life but his conscience, his self-respect, and his human dignity, the two form a single unit with all the advantages of man and machine acting in symbiosis.  They are in essence a hybrid creature, a cyborg.

And on Friday...

The day after making this presentation, I had an afterthought that shed new light on it. That notion will be posted here Friday. Don't worry, though. It's short.


No comments: