A lot has been written and editorialized about how ChatGPT is changing everything from education to journalism and even hiring. School districts are saying students are subverting their education by using it to write papers (read: plagiarize and cheat). Several articles have talked about how ChatGPT can take MBA exams and get medical licenses. That’s what it does when it has access to almost every written word and neural network capabilities to generate responses. I like how Cassie Kozyrkov explains it.
I have dabbled with ChatGPT. It’s interesting. I tried to get it to write a blog on a very specific topic. I wasn’t able to do it. I revised my approach to coming up with an outline. It produced a starting point for a table of contents that I might use at some point. I tried to have it generate social media posts. They were decent, but probably none that I would post without edits.
Like many other technologies at this stage, ChatGPT has promise, but it has limitations and its applications may be overstated. Many of the test cases being written about can be done today with Google searches. For example, If you are a student trying to find shortcuts, Cliff’s Notes and Wikipedia have been available for years (less the tools teachers have to sniff out plagiarized content)
Similar capabilities are being demonstrated in other areas of focus as well. AI tools are being used for both writing and reading resumes. Chatbots handle customer service inquiries. Even health systems have decision support tools to expedite treatment options.
However, I think about the evolution of these tools as “dimensions”, much like a 3-D experience. Early on, the engagement is flat, such as one or two dimensions. As we add more sophistication, as in a third dimension, we start to realize depth. A fourth dimension may be time. The same can be said for these technologies.
As the technologies evolve, so will their dimensions, moving into human experiences and space. To some degree, these tools will start to do things humans cannot, but they are not yet at a point where they replace a human’s ability to connect disparate themes with context and creativity. As humans, we should cede some elements of data collection, structure, and history to a tool. The volume of sources is growing exponentially and we cannot keep up with it.
Instead, we should interact with information and each other at higher levels of engagement. For example, what will you do with the facts once you have them? How do you apply them to the situation and make a decision based on real world parameters? How do your values, life experiences, and (yes) biases integrate into your process? Is there a technology that encompasses considerations?
