Recent news on artificial intelligence has raised furor among people. Many worry that jobs are going to be replaced by AIs and that humans will be left with no jobs in the future. However, the reality is not as dystopian as it sounds. The recent development of artificial intelligence is actually creating a great tool for us that help humans to become more productive than ever. But first, we need to understand how this technology works and how it applies to our daily life.
The "intelligence" in artificial intelligence refers to a mental process that takes place in human brains. Hence, what AI is actually trying to do is to replicate the steps of how human brains process information. In the process of acquiring knowledge, human brains take in information, identify patterns, and match these patterns to create a set of information that is called knowledge. Similar to what artificial intelligence is doing, the machines take in the huge amount of data, identify the pattern, and create a set of knowledge called analysis that provides insights that are useful for many kinds of fields. This process is called data analysis.
An important aspect that makes data analysis imperative in today's world is that it facilitates a process of what is called decision-making. Every problems posed in the world needs to be solved by making the right decision. Data analysis provides an essential input to decision-making process called prediction. Prediction is defined as processing information that we have to generate information that we do not have[1]. For example, we use the past inflation data to get data of tomorrow's inflation. In other words, prediction is the use of hindsight to gain foresight. These foresights help to assess the consequences of each decision, say in this case, to decide how much to invest in today's stock. By assessing these consequences, we will then be able to choose the most strategic and beneficial decision. Hence, the more accurate our prediction is, the better the decisions that we can make.
In order to make prediction more accurate, more amount of information is needed to be processed to generate patterns. With limited human brains capacity, analyzing huge amount of data would be an exhaustive task. In addition to that, as Daniel Kahneman argues, we as humans are noisy thinkers and we possess all kinds of cognitive biases[2], makes humans poor at predicting. With humans being poor predictors, it is therefore beneficial to use machines as predictor tool because not only that it can process huge amount of data, its analysis is also free from biases.
THE FUTURE OF WORK
That being said, advancement in predictor machines would mean more jobs are going to be replaced. As we know, this kind of revolution is not the first phenomenon that ever happened in human history. Along the history, we have experienced three types of industrial revolution. However, entering the fourth revolution, people do not seem to get any less concerned. In fact, they become more worried than ever. This is because the previous revolutions had only replaced administrative jobs and jobs that require heavy physical skills. As for today, new technology is going to replace higher level of human skill, such as data analysis, a task that has been long held by the middle classes. If we fail to acknowledge such changes, then those who have access to technology would have more control on the economy while those who do not will get even more left behind. Hence, it therefore raises a question, how can we, especially for future analysts, deal with this new change?
PREDICTION AND JUDGEMENT IN DECISION MAKING
Ajay Agrawal, along with Avi Goldfarb and Joshua Gans, recently authored a book called, "Prediction of Machines: The Simple Economics of Artificial Intelligence" to express their views on how to embrace the advancement of artificial intelligence using an approach to classical economic perspective. First, Agrawal argues that advancement in machine learning lowers the cost of predictive data analysis. As the price for analysis lowers, more businesses are going to use this tool to make faster and better decisions. The shift to the automation of data analysis causes the value of its substitute, which is human's analytical skill, to decrease. This is the part that causes worries among many because it replaces the job of conventional data analysts. However, what everyone fails to realize is that when the value of its substitute decreases, the value of the complement to data analysis increases. And in the decision-making process, the complement to data analysis is human judgment.
Human judgment is the ability to know which prediction or analysis is going to matter to a company. It determines how predictions are going to be used in executing certain decision[3]. To put it in analogy, prediction tool is like a new form of calculator. Calculators may generate numbers, but those numbers are meaningless if humans do not know what to do with the numbers. The same goes with prediction tool. However, instead of calculating numbers, it matches patterns and generate data and these data do not speak by themselves. Humans still need to translate the use of these data and use it where appropriate. Hence, even though data analysis will be automatically generated by machines, human judgment will just be just as important in business' decision-making process.
An example of collaboration of automatic technology and human judgment can be seen in financial economics sector. When artificial intelligence wasn't as developed, predictive modeling was done manually. Financial models that used to take days to process can now be processed in a short time and large scale. Prediction about trends of stock and trades can now be accessed timely by hedge fund securities and by investors.
In this case, automation plays role in accelerating the process of providing predictive analysis. However, the decision about which stocks to invest in still requires human judgment. Hedge funds securities need human judgment to assemble information and decide which information that should be included to make a good portfolio. The same goes on the other side where investors also need human judgment to decide which portfolio to invest in[4] .