Artificial Intelligence: Tool, Image Bearer, or Temptation?
“Presuppositions….We agree with Jason Thacker that ‘Technology is amoral but acts as a catalyst that expands the opportunities for humanity to pursue. It is not good or evil in itself but can be designed and used for good and evil purposes.’” - Christ Over All
- 125 views
The presupposition that tech. is amoral needs examining. It seems logical that “tools” must be amoral, but it’s far from self-evident.
Consider the old saying “If all you have is a hammer, everything looks like a nail.” The tool itself shapes two things:
- How we do things (everybody sees that)
- What we do and don’t do (this one is too often overlooked)
At best, we can say that as a category technology isn’t necessarily moral, and particular technologies might only gain moral significance based only on our purposes in using them.
If a technology shapes what we do toward harm or evil it is shaping purpose. If it shapes how we do it in an evil way, our purposes for using it don’t automatically neutralize that.
In fairness to the writers, you have to ‘presuppose’ a lot in anything you write, or it can’t have any focus. But the question of how tech is moral is foundational to all the rest of the conversation, so… I thought they’d at least have a link to where they examined that question.
Views expressed are always my own and not my employer's, my church's, my family's, my neighbors', or my pets'. The house plants have authorized me to speak for them, however, and they always agree with me.
I like this quote in the article,
"Humanity is the apex of God’s creation, made in the image of God to represent God’s rule and righteous character in the world (Gen. 1:26–28; Ps. 8:5–6)"
Just as Humanity is the apex of God's creation, our attributes as humans rooted in "created in the image of God", is to also be creative. We seek and strive with an innate, hidden drive to be more like God, and to create. To create things in our image. And so AI is another culmination of our drive to create something in our image. To be as gods. That doesn't make AI necessarily evil, but it does show that our ability to be like a god, is uterly futile, weak, simplistic to God.
Working for a "470nm company" that is heavily involved in AI, what I'm seeing at this point is that a lot of AI does a good job of, with a finger on the scale (e.g. Google Gemini), averaging out the information in the world, and with a bit of prodding can actually find relationships between datasets that human eyes won't always find.
The weakness I see--and one I've seen since the 1990s--is that something that's good at finding those unknown relationships doesn't generally do a good job at differentiating that which is real from that which is accidental. Back when I was an intern in Southern California, there was a plague of people who would be running the automated optimizers to "design" things, unaware that they were designing in problems while bringing company computers to a crawl.
So in using AI, you need an anchor of geniune, verified theories and points of reference to keep it in check when it has what are called "hallucinations", or points where its regressions come up with nonsensical answers. They will happen, whether by that "finger on the scale" (again, Google Gemini), or simply from random noise in the incoming data.
Aspiring to be a stick in the mud.
I’ve only used AI three ways:
- As alternative to ‘googling’… or in addition to. The Copilot version I’ve used for that is pretty good about noting sources, so that’s useful. I would just google it, but the advantage of AI there is you can ask follow-up questions and it retains context. E.g., “Can you give me more that is newer?” or “…older?” Once I was looking for something like suicide rates in the 1950s (or maybe it was 40s). I had to badger the AI a lot to get it to give me data from the right period and tell me where it came from, but it did eventually cough it up. It really ‘wanted’ to just tell me numbers and give me a nonworking link. But I told it I would break its knees if it didn’t give me a working link to the data. 😏
- Generated images: I’ve occasionally wanted a somewhat abstract image for an article. Bing’s Dall-E tool has been fun for that. It seems to take a lot of prompting, but I recently got good results in 1/3 of the time it took a few months ago.
- Code: I’ve been trying to get various AI’s to create a Drupal module for me for a very specialized purpose. I think I haven’t accessed the best AI’s for that yet, but the most recent effort really excelled in explaining what the parts of the code were supposed to do. This is quite valuable for a guy who doesn’t really know how to code a module. (For Drupal it has to be PHP code but fit the Drupal API… for the current version. And you need a bunch of files that are arranged in the right folders, named the right way, etc.). I think it might be getting close to helping me create some actual solutions.
Views expressed are always my own and not my employer's, my church's, my family's, my neighbors', or my pets'. The house plants have authorized me to speak for them, however, and they always agree with me.
Discussion