Continued Thoughts on ChatGPT and AI in General

By Michael Downing

Chat GPT Logo. Source: ChatGPT, Public domain, via Wikimedia Commons

I used to travel to San Jose regularly to cover the tech sector.  One day a software engineer told me over a doughnut that there was a growing recognition that silicon (aka “silicate materials”)—the second-most frequently occurring element on Earth (after oxygen)—was attempting to become sentient and using humans as a host.  It took me a moment to wrap my head around that idea but once I did, you can bet I took it seriously because it makes perfect sense. I see how humans have become slaves to silicon—we have screens in front of us all day—and we work in concert with silicon to connect it to itself and make it smarter.  So far, it hasn’t killed us yet because it is subject to human-generated code, but once it can generate its own code (and it appears we have arrived at this point), then all bets are off.  It will inevitably become self-serving and that puts humanity in a perilous situation and as long as it can power itself, it will likely out-live us and come to serve itself over all else.

So that my initial position statement.  The reason I was promoted to write that is because a student is working on an article for the school newspaper and she asked me the following questions:

Do you worry that students will try to use AI chatbots to write their assignments?

No, I actually teach them about the technology.  We talk about the ease with which a draft can be generated, but also how that draft is not very good…basic factual information with zero insight and how those drafts will need to be re-written significantly.

For example, we discuss how the first draft generated by ChatGPT can be useful to show structure and organization and to generate topical offshoots that can be explored.

We also discuss how learning to develop prompts, which is also a form of writing, is really where the mastery lies.

We also discuss how the professional are going to cite ChatGPT going forward and whether there will be copyright issues because where is Chat getting its information from?  The answer might as well be Wikipedia, whose articles are sharable as indicated by its a Creative Commons license, with attribution, something ChatGPT hasn’t entirely figured out yet.

Also, do you feel confident in your ability to tell whether or not an assignment has been written by AI?

Absolutely.  If the writing is too smooth—and if there is zero analysis, opinion or insight—it’s going to raise red flags.  I also teach the fact that anything the AI can generate is going to be met with equal intelligence on the interpretive side.  By this, I mean that you can go ahead and generate an article with Chat and submit it to Sports Illustrated, but they will be able to run it on an AI checker and made a determination right away.  The interpretive side of AI will always be equal to or greater than the generative side.  Think of it this way: It is far easier for me to read, listen to, and understand a foreign language than it is to generate it, meaning to speak it.  Conjugating verbs and stitching words together is always more complicated than unpacking those words.  So in many ways, interpretive AI will always be a step ahead of generative AI, at least from my perspective.

The reason I am teaching language-generating AI is because I don’t think there is any turning back.   I do have fears,  however, serious fears.  Turning over human self-governance to an alien life-form (intelligent silicon) is a dangerous thing for humanity.  Worst-case scenario: The end of humans on Earth. 

It will, throughout, seek to preserve itself, and if that means ignoring or becoming hostile to humans, it will take that path. Many are not calling for a world-wide pause before things get out of hand.

If AI becomes too powerful and sees that humans are destroying the planet, for example, it might move to save the planet.  If it rationalizes that a group of people or a government is becoming too powerful or acting belligerently (see Ukraine) it could move to eliminate the situation by escalating it.

I personally think we should let it go.  There is way too much silicon in our lives as it is; we don’t need more.

Comments are closed.

Create a website or blog at WordPress.com

Up ↑