Chat GPT — Planning for the Right Target
A Battle for Process Over Product in the Age of Chatbots
For Christmas, we tried to pare things down. Our kids each got to request one gift from Santa and we supplemented with a board game — for my son, an 8yo strategist, chess player, and self-proclaimed gamer (which is really just for an hour on Saturday and Sunday afternoons), we purchased an automated version of Battleship, where you can select a programmed layout, plug in your pieces, and immerse yourself into the sounds of sea combat.
For our first game, we decided to play as a family (2 v 2) and skip reading the instructions for manually inputting a layout; we selected a predetermined one, planted our battleships on the grid and pressed “go.”
As torpedos whizzed through the air, and we waited on baited breath to hear whether a ship had been hit…I realized something was missing.
We were so focused on the distorted crackling coming through the cheap plastic speaker that the grid itself became obsolete. Our spatial and numeric processing skills became irrelevant in this new “advanced” version of the game.
I made a similar, perhaps more troubling, observation when my son asked me to watch the dexterity and speed with which he popped balloons to level-up in an online French reading program (the instructions were to pop the balloons with the corresponding letter sound). Popping many balloons worked just as well as popping the right ones. And I thought about how technology, even that which is designed to support learning, can have a dangerously adverse effect.
Automation, in these instances, tended to “make things easier” by removing the thinking or the procedural steps required for an outcome. Product took centre stage while process faded into the background.
OpenAI’s ChatGPT seems to be following a similar trajectory. The platform creates human-like responses to prompts in seconds. It can do almost anything — write a paragraph in the voice of a fifth grader, create decodable texts based on a specific phonics scope and sequence, generate lessons, create dual language texts, provide feedback on student work using a specific rubric.
The responses are not always accurate and vary from time to time, but the system learns through feedback; it is iterative and responsive.
I think about Battleship, and my son’s ability to manipulate the results of his online reading program by bypassing “the thinking” for dexterity and speed, and I wonder whether ChatGPT will allow students and teachers to do the same — to bypass the process for product.
What will this new technology mean for education? Will it eliminate the painstaking yet crucial reflective process that often accompanies learning and teaching?
TO CONTINUE READING, CLICK ON THE PUBLICATION ON ELAN’S WEBSITE.
