Group 04

Chat, Write Me a Proposal on the Benefits and Harms of AI

Author
Affiliation

Byeong-Hak Choe

SUNY Geneseo

Published

November 25, 2025

Dear Group 4,

Thank you for your kick-off report on the benefits and harms of AI. You’ve chosen a very timely and ambitious topic and correctly centered environmental externalities (energy, water, and emissions) alongside potential benefits (efficiency, forecasting, innovation). The project’s potential for a strong final proposal is high if you (i) narrow the scope of your main question and (ii) anchor your analysis in a clear environmental economics framework.

You are working toward both the presentation and the final proposal. Focus on economic reasoning supported by existing literature and descriptive evidence.

Next Steps

  • Clarify your main question
    • Decide whether you are:
        1. Evaluating the net balance of harms and benefits of AI, or
        1. Focusing primarily on the environmental harms and how policy should respond.
    • Rewrite your main research question in 1–2 precise sentences.
  • Choose a small set of indicators from existing sources
    • Select 2–4 concrete indicators such as:
      • Data center electricity use
      • GHG emissions from data centers
      • Water use for cooling
      • One or two benefit channels (e.g., AI-enabled energy efficiency or better forecasts).
    • Use secondary sources only (reports, articles, etc.) to obtain these numbers.
  • Anchor the project in environmental economics concepts
    • Make the economic mechanisms explicit:
      • Negative externalities (energy, water, emissions, local impacts)
      • Dynamic efficiency (short-run innovation vs. long-run climate damages)
      • Possible rebound effects (e.g., more consumption because AI makes tasks easier/cheaper).
    • Organize harms and benefits within a simple cost–benefit narrative or marginal analysis over time.
  • Use visual evidence from existing sources
    • Consider including one or two simple visualizations in your slides (not in the proposal text) such as:
      • A visual or table comparing data center electricity use or emissions over time or across regions.
      • A simple figure showing projected growth in AI-related electricity demand.
    • These visuals should be based on published tables/figures (you can recreate a simplified version) and clearly cited.
    • This is about using existing descriptive/statistical information to motivate your question.
  • Draft policy implications grounded in economic reasoning
    • From your analysis of externalities and trade-offs, sketch 2–3 policy levers, for example:
      • Carbon pricing or tighter emissions standards for electricity generation used by data centers
      • Efficiency and water-use standards for data centers
      • Requirements or incentives for renewable procurement and locational choices
      • Disclosure requirements on AI-related energy and water use.
    • Make sure each recommendation logically follows from your analytical narrative.
  • Build on key references

Questions to Think About as You Refine Your Final Proposal

You do not need to answer all of these, but they may help you sharpen your narrative and recommendations:

1. Scope & Economic Mechanism

  • Are you asking, “Does AI bring more harm than benefit overall?” or “Given that AI is growing, how should we manage its environmental harms while capturing key benefits?”
  • What are the main externalities you want to highlight (electricity demand, emissions, water use, local noise/land impacts)?
  • How does dynamic efficiency enter your story? For example:
    • Short-run: rapid AI expansion, high energy use.
    • Long-run: AI may help optimize energy systems, forecast storms, or improve climate research.

2. Indicators & Evidence

  • Which 2–3 metrics from your sources will form the backbone of your environmental argument (e.g., projected data center electricity demand, estimated emissions, water withdrawals)?
  • Can you summarize these in a single table or figure for your presentation slides that clearly shows:
    • “AI-related activity is growing fast,” and
    • “This growth has non-trivial environmental consequences”?
  • How will you explain the uncertainty in these numbers (e.g., different scenarios in the IEA report)?

3. Governance, Policy, and Norms

  • What mix of policy instruments seems most appropriate from an environmental economics perspective?
    • Price-based tools (carbon pricing, water pricing, energy taxes)?
    • Quantity/standard-based tools (efficiency standards, cooling-water rules, siting restrictions)?
    • Information-based tools (disclosure, labeling of AI services’ energy use)?
  • How should we think about rebound effects (e.g., AI makes some processes cheaper → more usage → higher total energy demand)?
  • Under what conditions might AI innovation cross the line from socially beneficial to socially harmful in environmental terms?

4. Distribution, Location, and Equity

  • Where are data centers located, and who bears the external costs (e.g., local communities facing water stress, noise, land-use changes)?
  • How might your recommended policies address equity concerns—for example, between:
    • Tech firms and host communities, or
    • Regions that benefit from AI vs. regions that host the environmental burdens?

Best,
Byeong-Hak

Back to top