Skip to Main Content

PHIL 150: Philosophy of Technology (Spring 2025)

Generative AI

How it Works (*very* Broadly)

Pulling for a staggering number of sources, AI generates a human-like response. Factual data is pulled from the internet and other open sources. The tone and style of responses is modeled on reprogrammed exchanges & learned from other interactions with actual humans.

It is unclear where the data originates or which algorithms are used to generate the dynamic responses.

What it does well:

  • Synthesizing data. It pulls from a staggering number of sources.
  • Approximating human style. If you ask it to write like a pirate, it does a surprisingly good job of writing like a pirate.
  • summarizing information spread across many websites. (Example) 

What it does poorly:

  • Nuance. It will boldly present inaccurate facts as accurate.
  • Limited knowledge base. Something like 85% of academic works are behind a paywall, so AI is mostly limited to materials freely accessible online.
  • Bias. The analysis provided by AI is directly shaped by the data fed into the system. This can be a serious problem.

Ethics

  • Energy consumption:
    • 1 google search uses about 0.3 watts of electricity. ChatGPT uses 3
    • 1 Chat GPT search uses ~0.3 kWh ; that's the same as leaving an old 60-watt lightbulb on for 5 hours or a modern LED bulb lit for 30 hours.
    • By 2030, AI energy consumption is expected to increase by ~1.6X
  • Attribution:
    • the sources being used aren't credited.
    • who should be credited with the response?
  • Academic honesty:

Within the Context of Library and Information Science

Go back to our early conversation about Questions to Ask About Every Source

 

  • Who is the author?
    • Do they have an agenda
    • Are they an expert?
  • Where was it published?
    • Is it peer reviewed?
  • When was it published?
  • What is the argument?
  • Why was it created/written?
  • Is it trustworthy?
  • Is this source well-cited?

 

Compare an AI response to our databases again:

All of academic is designed around the idea of authority

  • authority of the author to provide their argument/assessment
  • authority of the sources they use
  • authority of the journal to circulate ideas

By it's nature, AI is unclear.

  • unclear where information is gathered
  • unclear who created the opinions addressed
  • unclear how information is gathered
  • allegedly even the programmers no longer fully understand the process. (ex)(ex)