Skip to Main Content

Ethical AI

Learn about how to ethically use AI tools in your schoolwork.

Navigating AI Policies

In every case, you will need to consider AI-related policies at multiple levels:

Institutional

(CSM and/or specific academic departments)

Implications of using AI based on CSM policies in the Student Handbook (including Academic Integrity).

Professor

Implications of using AI in a specific class based on policies in the syllabus or assignment directions.

Personal

What do you think? Do you want to use AI, or avoid it?

Consider how to advocate for yourself and protect your privacy, intellectual property, and environment..

Ethical Considerations & the Impact of AI

Privacy

  • AI can collect private or sensitive data, often without consent
  • Trained on existing digital content, often without consent
  • Unregulated use of AI (with potentially harmful biases) to analyze surveillance

Accessibility

  • Increase the digital and tech divide
  • Monetary and technological barrier to accessing AI tools and AI literacy (education)
  • Increase accessibility of written content through screen-readers and vocalizing text

Misinformation & Disinformation

  • Increase the production, spread, and use of false content (generally misinformation)
  • Misinformation = "the dissemination of false information"
  • Disinformation = " a form of propaganda involving the dissemination of false information with the deliberate intent to deceive or mislead"
  • AI Hallucinations are, at minimum, a form of misinformation
  • AI software is often inherently biased (trained on biased data)

Environmental Impact

  • Water usage: estimates of 500 ml or 16.9 oz per ChatGPT conversation (more for image generation)
  • Land usage: data processing centers require land and produce pollutants such as mercury and lead (leading to soil and air contamination)
  • Electricity: estimates of 2.9 wH per ChatGPT search, compared to 0.3 wH per Google search
  • Precious Minerals: technology requires use of scarce minerals, which are often environmentally destructive and dangerous to find and mine

Copyright / Intellectual Property

  • Legal gray area: new copyright laws explicitly state a work cannot be used to inform an LLM; generative AI tools are still fairly new and courts are still working out details on when and where copyright applies in relation to them.
  • Copyright = exclusive legal right to make or distribute copies, create derivative works based on the work, and publicly perform or display the work.
  • Alternatives to copyright protections:
    • public domain
    • open access
    • free to view
    • free to use
    • Creative Commons licensing
  • People and AI often ignore copyright or intellectual property rights:
    • Uploading content without the author's consent
    • Copying/pasting images protected by copyright
    • AI benefits (in money and upgrades) from learning and accessing new information, regardless of whether it does so "legally"