Telecom & Technology Consulting
By the time you read this sentence, chances are you’re getting a bit tired of hearing about the wonders AI can do for your organization. You experience those benefits every day: Vast increases in efficiency, exponential growth in productivity, liberation from much of the humdrum of modern work. But the legitimate rewards of AI can come at a price – one that erodes the very judgment and critical thinking that technical professionals depend on.
Despite the wins that generative AI creates, the technology simultaneously delivers much useless content. But IT pros have become so dependent on taking the path of least resistance, they believe that they’ll still produce the same high-quality results they did in the pre-AI era. This lazy approach, paradoxically, makes them less productive than before. Meanwhile, organizations see a reduction in quality control and, perhaps most crucially, integrity.
Tech leaders can course-correct, but they must act soon before control completely slips away. The problem they must eradicate is what some experts call “knowledge decay”.
To do so first requires verification, which means separating genuine human-created content from that produced by AI. As you can imagine, with the large volume of content generated by the average IT department, verification can take a considerable amount of time. This task necessitates the power of critical thinking and extensive research, two capabilities that are slowly eroding inside IT departments, mostly among younger technical pros. After all, if they’re used to getting the “right” answers simply by typing in prompts, they may not even have the actual ability to think critically and discern which research is legitimate.
Solving the problem of knowledge decay also requires knowledge validation. In this phase, you must confirm the true value humans have provided when using AI as a tool for a given task or project. As an example, a design firm could create a website using AI, when the client is hiring them primarily for their ability to solve user experience challenges.
This ethical dilemma extends to the technical services industry, both for external and internal clients. With the use of AI now ubiquitous, everyone assumes – and expects – that you’re using the technology to do your work. Nothing wrong with that, to a degree of course. The problem is that we’re inevitably asked how much work we did versus how much we relied on AI to do our thinking. In turn, we’re forced to justify our answers. Which is another reason tech leaders need to gain control of AI use.
Also consider that as content is put through AI repeatedly in order to hone results, it naturally falls farther away from its purity. And remember that LLMs are context-agnostic, meaning that they merely predict the most likely results of a given prompt.
In order to halt the downhill slide cause by overuse of AI in IT departments, experts call for a fundamental shift in how we carry out projects, and to concurrently develop rules for processes.
AI should strictly be used only in situations where it can add measurable value. To illustrate the rule, when creating content for performance evaluations, managers can gather detailed information from IT workers and clients, and then use their preferred AI tool to distill that content instead of having AI provide a list of generic bullet points.
Also assess how AI may affect or impact a specific process. For example, when looking at the revenue cycle, all key stakeholders should be aware, and agree on, the manner with which AI may be used in the overall process. This isn’t a matter of debating if AI is the right tool all the time – it’s a matter of considering how more human involvement could ultimately boost efficiency and quality.
We must remember that AI is a tool, and despite it’s immense power to speed up work, we can’t allow ourselves to be fully reliant on it. We can’t let it suppress the need for critical thinking. But when we create the right balance between human judgment and use of AI, we’ll get the results we want for the short- and long-term.
