For millennia, philosophers and religious leaders have pondered the true meaning of free choice. Within Jewish tradition, the belief is that we do have free choice to do right or wrong, but our choices are already known to Gd. The obvious and impossible-to-answer question is: if our actions can be known before we perform them, how can it be said that we have free choice to do them.
If a regular person living today could know in advance how people would act, such a person would effectively have the power of a god. There are a number of wonderful science-fiction stories/movies which play with this specific idea. As it turns out, this is no longer just an issue for philosophers, divinities and fiction writers.
In the following article, the author speaks of the predictability of voting choice based on other behaviors. Let me already say that such predictions are not 100% accurate. However, given time, improved algorithms and more and more information available online, the accuracy of such predictions will approach perfection. In terms of voting, one does not need to be 100% accurate. Excluding one movie that fictionalized a presidential vote that came down to a single person’s choice, it is definitely sufficient to be 95% accurate [or even less], in order to predetermine the outcome of an election.
Why is this important? Why does knowing the outcome of a vote help anyone? Actually, it is one of the most important pieces of information that people can obtain. If a party [in a democratic society] knows that a particular population in a particular region of a country, will vote a certain way, resources can be directed at this group or away from this group based on the desired outcome. So if the people of Ashkelon primarily vote for one party, that party can effectively ignore this population. Other parties, on the other hand, may choose to invest time and money in convincing the people of Ashkelon to change their decision. Each month, the same analysis could be done to see if there is a change in the predicted votes of the Ashkelon community. And based on a change, all of the parties could change their strategies.
To paraphrase Austin Powers, such power does not need to be used for “evilness”. Predicted purchasing behavior is already used by major stores to fill their shelves with the products that people will buy. If a company knows that there will be a tremendous demand for yogurt, the same company can make sure to always have a sufficient quantity in stock. Contrarily, the same company can repurpose shelf space previously used for “Milkies”, that apparently are now only purchased in Germany.
In the medical world, predicting behavior has tremendous value. For example, let’s say that a doctor has identified high blood pressure in a patient. Based on a data analysis done by a supercomputer in the NIH, the doctor can already know that this particular patient will not take the medication. It could be that the patient will not take the medication for reasons unrelated to the high blood pressure. Those reasons could be cost, or distance to the pharmacy or many others. The point is that to personalize medicine, which is a key goal in the future practice of medicine, one needs to know if patients will abide by physicians’ recommendations.
Let’s imagine a different scenario. A patient is told by the doctor that he must lose weight and exercise or else suffer a heart attack in the near future. Based on the same type of NIH analysis noted above, the doctor already knows that the patient will most likely start a full-scale exercise and diet program. If so, the physician can hold off on starting the medication, and instead reassess the patient in a month. This saves time, money and other resources for everyone.
Being able to predict behavior has endless benefits in every realm. The obvious argument against using such predictive tools is that it reduces people to players on a chessboard. People will begin to feel that they are trapped within a predestined path, where their choices are manipulated by powerful external forces.
When a person decides to purchase a particular item, they will wonder if this is a truly free choice, or if their actions were predicted. If upon entry to the store, the individual is already emailed a 20% off coupon for the item that they thought to purchase, it is clear that the choice was predicted. Interestingly, this simple fact might make the individual choose not to purchase the item, as if this could somehow foil the entire predictive algorithm. I will leave it to the reader to decide if such actions are effective.
Many children are acquainted with the phrase, often spoken by their mothers, that “I know you better than you know yourself”. I can speak from personal experience that such a phrase was incredibly upsetting. It was as if I was denied a level of self awareness. If someone else could know me better than I knew myself, how could I ever trust my own decisions. This will be the reaction of people to such predictive systems. But there are ways to avoid this reaction.
Unfortunately, most of the ways to avoid a negative reaction to predictability, is to simply hide the fact that the person’s behavior was predictable. In reference to the scenario of purchasing an item in a store, the owners of the store could avoid any uncomfortable feelings by sending an apparently random set of coupons to the email of the potential customer. Within the set of 3 to 5 coupons would be a coupon for the item that the customer is predicted to purchase.
We are creatures of habit. Even before super powerful modern computers and endless data streams from Twitter and Facebook, psychologists and behavioral scientists studied the nature of human action. It has long been recognized that humans tend to follow a pattern in their decisions and actions. The only difference today, is that we have the tools to quantify behavior to the extreme. No one has taken free choice from us. But our behaviors are a basic part of our personalities and ultimately define who we are.
Sudden changes in our behaviors can be an indication of disease, psychological and/or physical trauma, drug use, depression and more. Identifying acute changes in behavior, based on a deviation from predicted actions, may be the only way to spot a serious problem before it is too late. Certain changes in behavior could predict suicidal behavior and save countless lives. As usual, the same technology can have tremendous benefits but also bring with it significant risks.
I come to the end of this blog post and for those who are regular readers, you already know what I will write last. But it’s still my choice to do so.
Thanks for listening