How to Correct for Technology Bias in Intelligent Systems (2014)

 

What is technology?  The Museum of Science discusses and treats it by consensus as the tools around us that help us to work, learn, and live.   Technology specific to neuroscience is iconic: functional Magnetic Resonance Imaging (fMRI), electro-encephalogram (EEG), Microelectrode Single Unit Recording, Optical Imaging, and more.  These technologies allow scientists to observe what is going on where and when in the brain while the subject is undergoing specific designed tasks.  For example, an experiment may have a subject read the alphabet aloud as the researchers observe either changes in blood flow (e.g. fMRI) or ionic current (e.g. EEG) in the brain.  The researcher may even take the study to another level using another technology called Transcranial Magnetic Stimulation (TMS) to temporarily simulate a regional lesion.  Doing so may disrupt the subject in reading aloud the alphabet as if the patient had a temporary brain damage-inducing stroke.  This may be useful in understanding how individuals might suffer and potentially recover from brain damage.  The experimental subject by the way recovers perfectly within a few minutes and may suffer only minor local scalp muscle soreness if the TMS emitter needed to be repositioned and activated repeatedly to find the correct location.

 

But there is another, apparently under explored facet to technology.   Users use technology.  But the technology also uses the user. 

 

The scientists and researchers and engineers and explorers build different technology tools to fit their needs.  But the technology tools in turn alter what the scientists and researchers and engineers and explorers feel are their needs.  Let us trace through an example thought experiment.  Maps.

 

Throughout history, there have been two types of guidance maps: let us call them directional and overhead. A directional map is a list of directions tracing out a route.  It might consist of a list of text, images, or notes such as:

Driving directions to 1251 Avenue of the Americas, New York, NY 10020 
from 31 W 52nd St, New York, NY 10019:
 
1. Head southeast on W 52nd St toward 5th Ave
  
2. Take the 1st right onto 5th Ave
  
3. Turn right onto W 49th St
  
4. Turn right onto Avenue of the Americas
 

 

Which happens to be how one might drive between my old office locations.  Note how the instruction set is concise, detailed, and traces a singular route.  This is how one explorer might transmit knowledge of a previously traveled route to another.  These instructions are perfect for a single traveler – say a driver – to follow while simultaneously watching the road ahead for hazards.  The downside is that there are no provisions for exploring off the route or finding a detour should any portion of the route be blocked. 

 

In contrast, below is an overhead version of the same map.

 

A picture is worth a thousand words.  Unfortunately, in terms of size, readability, and cost to create, it certainly costs much more than a thousand times that of a directional map.  Pulling out this map, orienting it, locating it, and tracking on it is certainly not something for the single driver to be doing while traversing city traffic.

 

There is no doubt that technology is vitally important.  What remains constantly to be explored is how parallel technology choices affect our operations, our goals, and our ways of thinking and solving problems.  Quickly and conveniently give a friend the directional map over the phone and they may be at the destination quickly or not at all as they wander helplessly down the Lower East side since Fifth Avenue was blocked due to construction.  Hand deliver the overhead map to another friend – might as well personally guide them – and they may arrive in decent time with freedom of movement by re-routing around the construction.  But they may need to pay more in auto-insurance due to some fender-benders along the way.  Both maps are viable.  Both maps are vital.  The question might boil down to: how much effort can we provide and how much do we trust the environment and our friend’s capability?  Robust freedom and independence or robust guidance?  Readers in different countries may come to different conclusions.

 

It would be folly to believe that technology is truly linear.  There is no such thing as, “Country A is more technologically advanced than the savages at so-called Country B.”  The solution is not always to get more technology – it might be well nigh impossible if technology is non-linear since “more” is undefined.  Perhaps another way to discuss this non-linearity in one dimension is to discuss it in two dimensions.  How does one compare north vs east vs south vs west?  Then North Country has north technology whereas West Country has west technology.  But the connotation here is clearer that different technology tools serve different purposes.  A hammer is not more or less technology than a screwdriver.  But a user with a hammer would be biased towards hammering while a user with a screwdriver would be biased towards… it goes without saying.  Users operate the tools.  The tools also operate the users.

 

Our current technology relies heavily on computer hardware.   Our current neuroscience technology relies heavily on computerized scanning and location hardware.  These tools are ever increasing in spatial and temporal resolution to locate and track precisely where and when neurons are behaving differently when we perform various tasks.  The ultimate “Holy Grail” would be to find evidence that a single neuron or portion of a neuron is responsible for this activity or that thought and thus to calculate our ultimate storage capacity.  But is that our research and understanding driving us or is that our tools driving us?  What if the brain does not work that way?  What if it works in ways consistent with many tools, but some tools provide different answers than others?  We may believe that our tools are advancing linearly towards that day when we can track and map single neurons.  But perhaps we may then end up like the screwdriver’s target… it goes without saying.