Big social data obligation to tell stories requires coercion, big data invites reconsideration of torture

A recently released Senate majority-party report on unauthorized CIA interrogation techniques describe very disturbing interrogation techniques described as torture.   Outwardly, the intention was to a milder form of torture under medical supervision in an attempt to assure the punishments will not exceed the body’s ability to recover.

One very disturbing message of the report is that these techniques were not authorized and were performed deceitfully.   The report alleges that lower levels of the government performed actions that in this case happened to be appalling.  Even if the procedures were more acceptable, we should still be alarmed that the government-paid staff and contractors are performing unauthorized actions or decision makers neglecting their duties to provide oversight.   To me, this is an example of unaccountable actions occurring throughout the entire government.  This example gives vivid scenarios that should capture everyone’s attention, but the problem is probably more common than just the practice of torture.    In recent posts, I’ve been discussing my concerns about the lack of accountability in government where this can lead to an eventual withholding of super-majority consent to be governed.  The Senate majority-party report is this alleges that government staff and contractors are performing actions without leadership authorization or knowledge.  I object to these unauthorized and unaccountable actions of mid-to-low level government staff and contractors within bureaucracies even if their actions (or inaction) are more benign.

However, for this post, I’m going to assume that the appropriate upper leadership approved all of the actually implemented procedures, and that the lower-level staff kept leadership appropriately informed.   For most people the Senate-majority report provides alarming details of actual cruel methods employed as part of the enhanced interrogation or torture.

Speaking personally, I am disturbed by the idea of our government employing such cruel actions.   Despite my personal preferences, I think we should have the debate about whether our ideals that forbid such cruelty are in fact justified and realistic.   We take pride in our modern concepts of government, but we are humans with the same nature and life limitations as our ancestors.   Even though we find cruelty to be repulsive, cruelty may be a very realistic policy in context of human nature that has not changed recently.

In past posts, I discussed how the recent interest in big data promises are leading us to objectionable obligations.  I described how the volume, variety, and velocity of data can obligate a decision maker to follow machine-generated recommendations (through predictive or prescriptive data analytics).  I also described the need to maintain social harmony by obligating the population to cooperate with the decisions even though there is no accountable human decision maker.   Both of these are autocratic tendencies from the promise of big-data benefits.   With this recent discussion of the acceptability of cruel interrogations, we see a third obligation and that is the obligation to talk.   Demanding this obligation may require cruelty, but this cruelty is at least analogous to the cruelty that comes with the obligation to participate in decisions that will cause harm to an individual (such as what can happen in healthcare).

Much of the reporting has focused on the specific disturbing details of specific techniques actually used.  These are cruel procedures that the report authors suggest we should reject.   From my limited reading of the details, my immediate reaction is to agree that these should never have been done.   I tend to agree that our ideals of human rights should forbid such forms of cruelty.

Despite my concerns about the details of cruelty, I welcome the opening of the debate about whether our ideals may be too unrealistic.  This is consistent with some of my earlier posts where I considered that we may be wrong in condemning cruel punishments in other countries as human rights violations.   For example, it is reasonable to debate whether is it more human to impose long but humane prison sentences as an alternative to brief cruel punishments that nonetheless allow the punished to return to society sooner.   As an other example, medieval justice involving cruel punishments had a reasonable goal of maximizing the liberty of the punished.  In their time, that liberty permitted the punished to redeem his soul, but in modern times liberty will provide the benefit of allowing a person to rejoin society and contribute.

A complaint about medieval cruelty is that their poorer understanding of human physiology and psychology resulted in excessive punishments, and in particular permanent debilitating damage.  Today, we have better understanding of physiology and psychology of pain to inflict cruelty with consistent levels of pain and with less risk of permanent accidental injuries.   Modern practices of cruelty can be more likely to be recoverable without permanent disability.

I am trying to avoid the word torture to describe cruel actions, although my inclination is to use that term.  I avoid the term because it is not very well defined.   Torture includes techniques that can cause disfigurement, disability, or even death.  These consequences may either be accidental or deliberate depending on the incompetence or sadism of the torturer.  Of course torture also includes punishments that do not cause permanent damage or death, but I prefer to use the term cruelty in an attempt to distinguish the modern capabilities of at least making a more educated attempt to avoid permanent consequences.

Cruelty involves two elements: a human actor that inflicts the punishment, and a punishment that inflicts pain, panic, disgust, or emotional responses that the punished person reasonably would prefer to avoid.   In nature, we are all subject to the risk of encounter exactly the same painful experiences through no fault of any humans.   Every time I hear of some news of an horrific accident or natural catastrophe involving human casualties, I feel horror at the thought of what those casualties experienced.   Cruelty can result in similar experiences but they are even more objectionable when there is an human who either is applying the punishment or is in a position to stop it.   The human agent inflicting pain on another human, rather than the pain itself, is what I mean by cruelty.  This cruelty is a milder form torture that does not intend to inflict permanent injury or damage.

As an aside. I wonder if the modern access to autonomous machinery can eliminate the distinction between natural and cruel causes of pain.   Being placed in a robotic torture device is not much different than being in a car at the moment of a collision or trapped in a flood.   The painful consequences can be comparable and the actual action is not in control of another human.  Some medieval torture devices such as gibbeting worked similarly where the actual punishment occurs significantly later than the human action to place the victim in some chamber.  The torture occurs as a result of the machinery that traps the individual instead of specific actions of another human.   The actual infliction of pain and suffering is a consequence of the circumstances.   Similar circumstance can occur in nature, but we reserve the condemnation of cruelty to when a human is actively inflicting the pain on another.

In earlier posts (mentioned above), I discussed the use of cruelty as an alternative to imprisonment for punishing criminals.  The Senate-majority report on torture raises a separate discussion of the use of cruelty to compel story telling.  As I mentioned above, our need for obtaining data may lead to need to obligate people to tell their stories.  To get a more complete data set of stories, we may need to compel people to speak when they would prefer to keep silent.

The right to remain silent is a natural right.  All written, verbal, or gestural communication is voluntary.  In modern usage, the right to remain silent is a constraint on government from compelling testimony.   Even if the government attempts to coerce a testimony, the nature of communication is such that every person remains in control over whether and what to communicate.  Communication is impossible if the person does not allow himself to make the conscious effort to communicate.  Excluding the possibility of unconscious communication through unconscious body language or physiological responses such those measured in polygraph testing, communication requires willful conscious effort from the story-teller.

In social media, we observe many freely volunteered stories.  Such stories may be very short messages such as indicating likes of an article, short messages, or replies to other messages.   They may be lengthier communications such as this blog post.   There is a minority of the population who will freely offer new communications and even these may be limited to the stories they feel comfortable sharing.

For more reluctant story tellers, there is the technique of building rapport that offers a trusted environment to share stories.   Investigative Journalism often uses this approach to generate material for stories.  Recent journalism-reported allegations of rape and police misconduct provide evidence that these stories can be very unreliable.  The rapport-building approach does promise more abundant stories, but these stories individually may be as unreliable as coerced story-telling.

From a data-science analytics perspective, the intelligence is rarely gained from a single data point that tells us what happens.  This is contrary to the frequently promoted ideal of big data that if we had measurements of everything, than anything we have a question about will be captured in a measurement somewhere.   In my experience, this does not work very often.  Even when a direct relevant measurement does exist there is usually many reasons to discount its reliability.   Instead, I learned that we more often gain intelligence (discover hypothesis) by finding patterns in larger data sets that may not even include the specific case we are seeking.  Analytics of dirty data (equivalent of gossip or fable-telling) can give us new understanding about the world and suggest new avenues of inquiry.

Coerced story telling (through application of cruelty) may indeed be unreliable on an individual basis, but all story telling suffers this problem to some degree.  In addition, many people may have the fortitude to protect the most sensitive secrets no matter how severe the cruelty.

People may provide fabricated information as a means to be relieved of the cruelty.   Even if the narrator remains successful at protecting his secrets, the narrator still must construct an intelligible communication.  At a minimum, he must use sentences with subjects, verbs, and objects that make some sense.   The narrator will have to fill in reasonable words to produce intelligible sentences.   More broadly, he will have tie sentences together into a broader story that is credible.  To construct even a fantasy, the narrator will draw upon his education and experience to provide material to populate his statements so that his story is comprehensible and credible.   Those supplementary elements will expose some aspects of the narrator’s knowledge.   Even if he does not reveal the sensitive secrets, he can reveal important place names or potentially relevant bystanders who can provide more information.

We can learn intelligence from coerced discussions by comparing the stories from multiple sources that may offer their stories voluntarily, through rapport, or through interrogations.  The gained intelligence may not be an immediate answer to the desired question.  Instead, the answer may emerge through the patter of elements or parts of the stories.   These parts may suggest a pattern that can predict the desired answer, or a pattern that can suggest new areas of inquiry.

Data science works the same way with social media.  We can find patterns that can provide predictive powers based on the available data even if the singular data points may be individually unreliable.   The value of big data is that it is big enough to allow exploration of multiple dimensions of the data to expose truth that would not be revealed in any particular data point.   To enjoy this benefit, we need a lot of data.   Many big data projects exploit social media for volunteered story telling.  Voluntary story tellers remain a minority of the population.   In early posts, I suggested that there is value for journalists to obtain stories from reluctant story tellers even if their stories are unreliable.  We can obtain useful information from unreliable stories when we combine the content of stories from a large number of story-telling peers.

There will remain stories that are out of reach of voluntary stories from social media or from the result journalistic investigative practices.  More aggressive interrogation, including cruel techniques, can obtain new information from reluctant story-tellers.   From a data-collection perspective, such interrogations offers an additional data source to provide a more complete library of stories so we can compare and contrast with other stories.

Within a big data context, we need to obtain a more complete picture of current stories within a population in order to provide the opportunity to discover new hypothesis by comparing and contrasting different stories or story-elements.  Relying only on voluntary story-telling or rapport-based journalism is not sufficient.   Stories will remain that people will strongly protect as secrets.  Part of that protection is to avoid talking at all.   Coercion can compel them to talk and even if they succeed in protecting their secrets, their attempts to construct a compelling fabrication will require supplying credible details drawn from their experiences or education.   The individual stories and their elements may be very unreliable data, but when combined we may observe useful patterns to suggest new hypotheses that we can test by seeking out new sources of information.


2 thoughts on “Big social data obligation to tell stories requires coercion, big data invites reconsideration of torture

  1. Pingback: Economy of compensated opinions in a dedomenocracy | kenneumeister

  2. Pingback: Big social data obligation to tell stories requires coercion, big data invites reconsideration of torture | Hypothesis Discovery

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s