Algorithms - They Rule us

We give little thought to the underlying effects that rapid technological change is bringing while going about our daily lives. I want to bring to the reader’s attention of our use of technology which has a huge influence over our lives.  Today there is nothing that we do, no action we take which does not make use of the electronics and software. Whether we surf the net or book a ticket on-line or even drive our car, all the technology we use is controlled by algorithms.  Algorithms are a series of instructions which allow you to obtain a desired result.

Algorithms came to public attention after computers came into our lives. Actually we never stop to think that nature runs on algorithms. The movement of planets, the growth of plants and animals, the change of seasons and even the human body functions are all controlled by embedded programs or algorithms.

In our modern world all the electronic gadgets that we use today have embedded software (algorithms ) that make them run the way we expect them to.  The embedded software is all pervasive. It has touched all facets of our lives. There is however debate if this pervasiveness is desirable. Or is there a downside to it. In 2005 in the city of Memphis in the US  the police made so many arrests that the police department ran out of vehicles to transport the detainees. The reason was that the police had started using a new software which helped them do predictive policing. They overlaid available crime data with social housing maps and even outside temperatures. And then instructed the algorithms to search for correlations. Then focused patrolling was done in the identified areas. The result was that the number of arrests increased because the police were at the right place at the right time. This is just one example how algorithms are  helping societies identify and overcome problems. From dating websites, on-line share trading, retailing algorithms are determining how societies are run.

But there is another side to it.  Massive amounts  of data are collected by security agencies of all countries. They trawl the internet, telecommunications, social media sites etc. and gather a colossal amount of data. The only way to process this data is to use algorithms. The question then arises how this data is used and what parameters are set in the algorithms. If one is looking for say,  a terrorist. There is very little accuracy in setting the rules. For example a terrorist is a rare entity if we consider a city. If the parameters are too tight or narrow the target will escape. If the algorithmic net is cast too wide a lot of innocent people are going to be dragged into it.  When handling such large data there is no way to accurately pinpoint your target.  The concern here is a lot of people will suffer maybe an intrusion of their privacy or even an investigation just because some algorithmic parameters were tweaked.

‘Mistakes ‘ can and will happen. In 2010 the Dow Jones industrial average fell 1000 points in just a few minutes. This was later blamed on the high frequency trading algorithms which are used to do automatic trading. It seems the algorithms were competing among themselves to lower the share prices!

Algorithms which are used in scientific research eg. in climatology, drug design or medicine are very accurate. Any bug or flaw can be clearly pointed out looking at the output. Its in the social or financial field where errors are more difficult to spot as the above Dow Jones example illustrates.  Big corporations in the consumer goods space are now using big data and algorithms to pinpoint potential customers and favoring the ones already in the bag. This has the potential to create discriminatory business practices.  Recording companies are now developing algorithms which predict which songs or albums will be hits and which will be mediocre. AI and Machine Learning is taking over everything.

Today much of the big data which is gathered is held in databanks with the government or corporations. We are still at a loss how to use all that stored information for the benefit of the society we live in . But we have to guard ourselves against the next step which may be already upon us. That is predictive use of the data using algorithms. That has the potential to cause some real problems. If data is not used ethically or governed by proper rules, our future will have to be predicted with some pessimism.