Have you ever clicked on an Ad and bought something that you saw in your Social Media Newsfeed? Or, watched a video on YouTube just because it was in the recommended videos section? Have you ever wondered how you are shown content that appeals to you?
Our decisions are being shaped subtly, and we do not pay much attention to them. Every day we make many decisions based on algorithms used to predict our behavior, and we are shown content based on our activity.
With all this custom content and intelligent algorithms powered by large user Data, a very compelling question that often comes to our minds is:
If we’ll come to a point where algorithms would start harming our decision-making skills, making us too dependent on them?
So, let’s see.
Where’s The Fear Coming From?
Over the years, we have seen that our suggestions and Ads have significantly improved, and the content that we see is very accurate based on our likings and preferences. Have you ever wondered how all this works and how companies like Google and Facebook are cashing our clickstream activity and internet usage data to show Ads?
In this article, we will shed light on such factors and look into how modern-day Machine Learning Algorithms and Accurate Data Analytics play a huge role in shaping our everyday Decisions and choices.
As statistics suggest, the big data and analytics market will be worth $103 billion by 2023.
One may argue that these Algorithms based on our behavior make our lives much easier and are there for our benefit. But based on many experiments and detailed studies cited in the research paper, these intelligent algorithms are affecting our sense of making decisions.
Some researchers also worry that in the future, we might even lose our ability to make decisions.
Will Algorithms Erode our Decision-making skills?
A simple answer to this question would be a no; algorithms do not erode our decision-making skills primarily because they are developed to help us make decisions, not handling them altogether. In the end, it is always on us, the humans, to make the appropriate decisions. Moreover, no matter how advanced technology gets, algorithms should always be used with caution, and one should never rely too much on them.
Further in this article, I’ll mention some examples to make it easy for you to understand how algorithms affect us and our decision-making skills. I’ll also shed light on some important experiments and facts which will help you have an in-depth understanding.
Highly Recommended Articles:
5 Reasons Why Data Analytics is Important in Problem-Solving
Decision-Making Environments: Types and the riskiest type
Decision-Making Models Explained in Plain Terms
How Are Algorithms Making Decisions for Us?
One would ponder how these Machine Learning Algorithms work?
The fuel for these decision-making algorithms is data.
“Every day, three times per second, we produce the equivalent of the amount of data that the Library of Congress has in its entire print collection.”
Forbes
Now we know that an average user produces an enormous amount of data every day, and it is being recorded and saved at some database on the other side of the world.
The next thing to ponder is how this Data is used to create powerful insights and highly accurate recommendations?
Data Scientists and Machine Learning Engineers, and researchers who have spent years and years dealing with different kinds of Data and databases develop such algorithms that can derive useful insights and results from the big data.
A famous Entrepreneur, Aaron Koblin, in the field of Data Science, says that we can derive useful results even with a ridiculously complex and enormous amount of Data if we use the right tools and methodology.
The main steps involved in Data Science are narrowed down to the few listed below:
- The first step is to collect quality data that can be used to derive results. Data is the essence of data science. More quality data means better analysis and accurate results! The tools used to collect and store data are MySQL, Apache Hadoop, MongoDB, etc.
- Then comes the phase in which the pre-processing of this Data is done. The raw data collected in the first step is unfit for use and needs to be cleaned before it can be processed. This step is called the Data pre-processing step. Tools required for this depend on the kind of pre-processing and can be Python, Pandas, Spark, etc.
- The third step is to explore the dataset that we have created using libraries and tools such as Numpy, Matplotlib, Pandas, or Scipy. This step mainly involves finding out hidden patterns and features in the Data that can be used further for analysis.
- After carefully completing the previous steps, this step involves the Modelling of Data. This mainly includes building Machine Learning models to analyze the trends and patterns in data. The model can then be used for predictions and analysis. Some of the tools required for this step are Sci-kit Learn, Pytorch and Tensorflow, etc.
- The final and the fifth step in this Data Science pipeline is the Interpretation of Data or the deployment of the model. This includes the presentation of the findings and the insights gained from the previous steps.
The step involves answering business-level questions and providing answers based on the predictive and prescriptive analysis. The most critical skill required for this is the art of storytelling, curiosity, and creativity. The tools and libraries used to build interactive dashboards and presentations of the findings are Matplotlib, GGPlot, Seaborn, Tableau, PowerBI, etc.
How These Decision-Making Algorithms Affect Us?
With time, more and more people are relying on Algorithms to make decisions for them. If we examine our daily activities, we can see that these algorithms are affecting our lives more than we think. We are going to talk about some of these examples in this blog:
- Social Media:
According to a study, an average human spends around 150 minutes on Social Media in a day!
With so much time on Social Media Apps like Facebook, Instagram, Twitter, WhatsApp, Snapchat, etc., every user generates an enormous amount of clickstream data. This Data, including searches, site visits, shopping history, and streaming data, just to name a few factors, is used by the tech giant companies like Google and Facebook to show Ads to each user based on his/her data.
The Advertisements that we see in our newsfeed enables us to buy things online that we would not have otherwise. These custom advertisements are helpful for some, whereas others find it petrifying that their every move is being closely watched and examined.
- Shopping:
Similarly, in online shopping, we are recommended things that we are more likely to buy based on our location, history, and other tiny details. These suggestions, ratings, and eye-catching advertisements cloud our decision-making skills, and we rely on these algorithms to choose the best product for us. Most people find this quite helpful.
If we take a look at our surroundings, we see that more and more people are getting comfortable buying products online. This is an excellent example to demonstrate how our decisions and thought process has changed over the years, people who didn’t trust online shopping are now also hooked to sites like Amazon and eBay, etc.
- Dating:
This example clearly shows how our decision making is changing based on Algorithms, i-e, Dating Apps and Social Media profiles in this case. There was a time when people used to date based on real-life interactions and personality. Nowadays, people date based on Social Media presence and profile. In most cases, people in real life are quite the opposite of what they seem on Social Media. This is one of the main reasons why most of the online dates are not successful.
These matchmaking algorithms work by comparing which two personalities go together, and based on data about each person, the algorithm makes matches.
It is interesting to point out that people that use such applications often rely on the algorithm because they believe the numbers. This is how our decision-making skills in dating and relationships are affected by Algorithms.
- Content:
We often open a video and watch many more things than what we have planned. We get lost due to the Pop-ups, Ads, and recommended content and end up surfing content far different than what we had planned. This suggested and targeted content clouds our decision-making capabilities, and we often drift away from the content we are interested in.
- Self-Driving Cars:
From helping humans make small decisions to have complete control over our commute, algorithms have come a long way!
With today’s cutting-edge technology and super-efficient Machine Learning algorithms, we can travel from one place to another with just voice control or feeding the system with the drop-off location.
The Algorithms behind self-driving cars take care of all the decisions that a human usually makes. For example, they manage the route, steering control, speed, and everything else required for driving a car.
Self-driving cars are an excellent example of how Algorithms are taking over our decision-making capability. Apart from autonomous vehicles, maps and routing algorithms also choose an optimal route for us. Instead of thinking of a suitable path ourselves and knowing the directions, we just focus on the directions provided by the algorithm. Google Maps is an example of such routing algorithms.
Are Algorithms Better Than Us in Decision-Making?
With all the talk about algorithms making decisions for us and crippling our natural decision-making capability, a question arises: Are algorithms better at making decisions than us?
The topic is very controversial, and some people may believe that there’s no way computers can surpass our decision-making skills. Meanwhile, more and more research is being carried out in Artificial Intelligence and Data Science domains.
The huge increase in Data and the ability to accurately use this Data has made algorithms powerful, and in many cases, these algorithms are better at making decisions than humans.
Suppose we look at the real-life scenario of how the algorithm is much better at diagnosing diseases from X-ray scans. Algorithms are known for outperforming a whole group of doctors in such cases.
There are numerous other examples in which the decisions made by algorithms are very accurate. Algorithms are also better at making predictions as compared to humans.
Do People Prefer Algorithms Over Human Judgement?
A study done over a set of stimuli shows that people do appreciate the results of the algorithm over the judgment that came from people. The final sample included 286 participants (129 men, 157 women) of age 37, and they were assigned to study similar descriptions of a heterosexual woman and man. They were further asked the following questions based on their perception of the study:
- How attractive do they think the man would find the woman?
- How funny do they think the man would find the woman?
- How much do they think the man would enjoy a date with this woman?
Participants were advised about each of the questions from an algorithm and 47 people (from another study), and then their final decisions were recorded in a survey.
The participants then calculated the appeal, humor, and enjoyment of the date based on the advice given.
The survey results showed that the humans preferred algorithms decision over the decision perceived by the 48 people. They even shuffled the titles just to test their selection between the two, and the results were still in favor of algorithms, which shows how blindly we humans trust them.
Conclusion:
So, the final question that one may ask is that should we be worried about algorithms making decisions for us? The answer to this is straightforward and predictable; Algorithms and complex systems are made for the sole purpose of making our life easier and better, and hence there is no need to worry about it. If anything, we should combine them with our judgments to make the best decisions.
With that being said, we should also focus on our mental health and make sure that we use technology and algorithms carefully and do not let these algorithms take over our thoughts, but rather use these algorithms to refine our decision-making.
