The prominence of web 2.0 technologies in the past decade has led to a huge demand for virtualized services. Additionally, the ubiquitous online resources, made available by the cloud computing model, have not only multiplied opportunities for businesses, but also made them easily accessible. Furthermore, business strategists, together with information technology enthusiasts, have over time praised decision support systems as the main players in innovation. This evolution has allowed the development of interactive website designs leading to an upsurge of content management systems and social networking websites in the past decade.
These systems consume and consequently produce a lot of data that is golden in decision making if well processed and analyzed. Such data is in a wide variety, volume and velocity and is referred to as big data. People with the skills required for big data analysis are referred to as data scientists. Data science as an area of study used to be a dilemma in the past and still is to some organizations. This is so because of its requirement for all-round professionals who are skilled in variety of fields, and who can be trusted with an organization’s data. Additionally, a data scientist must not only be interested in the analysis and interpretation of data, they must also be adept in developing data processing and analysis systems.
In cases where third party systems are used, knowledge of the programming languages supported by those systems may be needed. Given the vast skills expected of a data scientist, businesses are quite often forced to train their staff on additional skills. The foundation of such trainings would be the ability to harvest data from multiple sources and prepare it for the intended use. If this training is not well done, user’s data may be compromised leading to erratic and biased decisions.
Moreover, in their quest to achieve business objectives, data scientist may ignore the need for personal data privacy. This is a major concern in the 21st century due to the capabilities presented by modern data systems – especially those that can be trained through artificial intelligence techniques such as machine learning and deep learning.
Unlike human beings who can differentiate between what is wrong and what is right about the data, machines may not achieve this with accuracy. For instance, when artificial intelligence is employed in recruitment, machines can learn which gender, school etc… Produces the best candidates and give candidates of a certain gender or school a priority when making future decisions. This can lead to discrimination. There are many examples of bias in AI systems. Comprehensive studies involving different sets of test data may need to be undertaken to compare machine decisions against those made by humans. In a nutshell, despite many years of effort to fight discrimination, technology may take us back many years back; especially if there is no way for human beings to monitor decisions made by machines.
Leave a Reply