What I'm including below is a response I typed out in response to a newsletter from "The Rational Optimist" that was forwarded to me. Looking at the name now, I think there is a good critique that I can write about the ontology of what such a name proposes. They very much subscribe to a techno-rationalist paradigm, but more on that at another time. Anyway, here's the email I dashed off in reply...
I'm very worried about what all of this generative AI is doing to people. More and more people are losing the capacity to think. I've been worried about this in my own life just with Internet search engines. It is good for our brains and bodies when things are kind of hard to do. All of these AI boosters don't look at studies that show cognitive decline in those who uses generative AI chat bots to "help" them write something. It's cognitive offloading to the point it is creating atrophy of the brain. The more someone uses generative AI to do something the less neural activity there is and the less chance they'll have of remembering anything of what they "learned." And then thinking students are going to use it for the best of its uses is laughable because students are always going to try and take the shortcut.
Here's an article about the study from MIT that shows the cognitive debt from GenAI use: https://time.com/7295195/ai-chatgpt-google-learning-school/
Being told how to do something is a lot different than figuring out how to do something on your own. Actually reading a book is a different experience than reading a summary of the book. Listening to interviews with Warren Buffett and trying to internalize his way of thinking about things is different than asking a GenAI to talk like him. One thing is using your brain like a muscle and working to improve its capacity to think and the other is skipping the whole thing. People saying they are doing all these things with GenAI is so empty to me.
If someone claimed they walked a mile while just standing on one of those moving sidewalks at the airport, everyone would look at them like they were silly. But when a person prompts a chat bot to write them an email and they say they're getting so much more work done, they aren't. They aren't actually doing any of the work. Sure the person on the walkway moved a mile down the way, but they did nothing for their personal improvement. All this GenAI chat pot nonsense is claiming it's better for all the output while missing that the person who prompts the thing is losing the ability to think.
All of the kids of tech firms are in low to no-tech households. Most tech people will talk about how powerfully dangerous their products are and make sure their children are protected from it's harms. All this GenAI stuff is going to do is continue to entrench the difference between the privileged and the not privileged. The person who can afford the tutor will get the tutor and the person who can't will get the crappy AI tutor that barely makes a difference. When an article talks about the benefits of using an AI tutor it's when it's in conjunction with a person. Here's something from Brookings about that: https://www.brookings.edu/articles/what-the-research-shows-about-generative-ai-in-tutoring/
There are just so many ways to go wrong with the use of GenAI in the workplace, the school, and society it does not seem worth trying to use it for anything. There are so many necessary guardrails and safeguards to prevent from causing real harm that it's use cannot be justified right now.
Sorry for the rant, but there's just so much wrapped up in GenAI, and I have been doing a lot of reading on technology recently. Plus I've been analyzing my transcripts on my dissertation and over and over again students talk about the most meaningful part of school being the relations they develop with their peers and teachers.
Comments
Post a Comment