An algorithm is a step by step method of solving a problem. It is commonly used for data processing, calculation and other related computer and mathematical operations. [Technopedia]
Algorithms are becoming the master of our fates. An algorithm decides the results of our Google search, whether we are entitled to receive a loan, the price of our health insurance, or the likelihood we commit a new crime.
Our behaviour is tracked, predicted and influenced by algorithms. An algorithm collects, classifies, structures, aggregates and analyses all the information made available to it and takes its decision. The majority of us suffer the consequence of an algorithm’s decision with little or no questioning. We tend to believe that algorithms are fair, right, and unbiased. They are mathematical models after all.
However, algorithms have built in human errors, as well as conscious and unconscious biases. Algorithms can be designed to deliberately deceive regulators (e.g. emissions controls, traffic managements or price-fixing). Or they can become a means of “propaganda.” Lastly, algorithms can become too complicated for humans to understand or unpick (Andrews, 2017).
Hence, automated decisions making can result in: (i) loss of opportunities; (ii) economic loss; (iii) social detriment; and (iv) loss of liberty [here].
How we ensure that algorithms are designed to achieve the greater good, instead of resulting in harmful outcomes is an issue we have yet to find an answer to.
How the algorithm is encoded, how it is trained and managed it is the beginning of the story. Addressing human behaviour is only one part of the problem. How the algorithm is deployed, managed and governed by corporations is the obvious consequent issue. And last but not least, how the algorithm develops, through machine learning is to be considered (Andrews, 2017).
An algorithm is capable to result in an infringement of consumer protection, privacy and competition rules, among others. The assessment of the societal impact of these new technologies calls for a multidisciplinary approach that brings together economists, lawyers, experts of computer security and artificial intelligence as well as philosophers (here).
[To be continued]