by Carl James | BLOOMINGTON, IN | Nov. 23, 2023
I.T. support now needs to embrace data work because that is where technology users are. We in this field have long had an attitude of get the computer working for the user on the organizational network and leave the rest to either dedicated trainers or the internet. I feel it is becoming more and more important for white-glove style I.T. support to become knowledgeable and provide more help for how user are going to access and analyze data.
Last month after one year of studies I earned a masters degree in data analytics. Despite a long career working with technology and a CompTIA A+ certification, I still lacked a major I.T. degree. I chose data analytics because I knew the program would let me do the bulk of the work in Python. Western Governors University degrees are "self-paced", which boils down to paying for how much time it takes (number of six-month semesters). Choosing a program using technology I was already familiar with allowed me to complete the degree with minimal costs.
I strongly believe this was valuable whether I pursue a career directly in a data role or not. First, there isn't a knowledge job out there today that can't be helped by better understanding data. Second, I learned in a very deep way, how important good quality data inputs are to getting models that produce truthful insights. This is so important as the formal data roles are not usually the ones collecting the data to begin with. Third, supporting people who collect or perform data analysis is easier to do if the support people understand what it is these people are trying to accomplish.
It is no longer enough to get a knowledge worker online and setup with organization email and let them go. The users are now demanding more from the support team and if human to human support is to remain viable, the support teams need to level up their skill sets. Computers have been at the center of every office for over a quarter of a century now. Most new workers can't remember a time without easy access to the internet. We must adapt and grow with what our users are needing.
My studies also included sentiment analysis with natural language processing (NLP). Open AI's ChatGPT has thrown NLP onto the forefront of common knowledge. What we understand as artificial intelligence today is largely computers interpreting patterns from vast amounts of data. NLP such as ChatGPT allows for AI systems to interact with humans. Our role as support technicians should include guiding our users to understand what the possibilities and limitations of NLP tools are.
The results of NLP or any other modern AI tool are entirely dependent on the quality of the data that is fed into the models. For ChatGPT to simulate human language well, it has to have quality examples to go by. There is also a distinction between seeming competent and providing factual information. A model can provide one without the other. It is our responsibility to help users know how to use the tools to validate the truth of an A.I. response. For example, Microsoft's Bing Chat provides citations for it's A.I. responses to allow the user to validate that the A.I. understood the context of the original source. I can say from experience that NLP A.I.s are capable of grossly misinterpreting the source information.