AI-101

Lesson 5

AI Ethics and Critical Thinking

AI-generated

Learning Objectives
  • Understand AI bias and where it comes from
  • Know the basics of AI privacy: what happens to your data
  • Recognize the environmental cost of AI
  • Think critically about AI-generated content and misinformation
  • Make informed choices about when and how to use AI
Introduction

You now understand what AI is, how it works, and where it fails. Before we dive into practical skills, we need to address a question that does not have a simple answer: should you use AI, and if so, how?

AI raises genuine ethical questions:

  • It can perpetuate biases
  • It consumes significant energy
  • It can be used to generate misinformation
  • Your conversations might train future models

These are not reasons to avoid AI entirely. But they are reasons to use it thoughtfully. This lesson gives you context to make informed choices.

Bias in AI: Trained on Human Data, Including Human Biases

AI learns patterns from training data. That data reflects human society, with all its biases.

How Bias Manifests

Representation gaps:

  • If training data has more text by/about certain groups, AI knows more about those groups
  • It might default to Western cultural contexts
  • It might perform better on tasks related to well-represented groups

Stereotype reproduction:

  • AI might assume a doctor is male and a nurse is female (unless specified)
  • Historical inequities in the data become "normal patterns" to AI

What AI Companies Are Doing

  • Curating training data
  • Training models to refuse stereotyped outputs
  • Testing for disparate performance across groups

These efforts help but do not eliminate the problem.

What You Can Do

  • Be aware that AI might reflect biases, especially on sensitive topics
  • Question assumptions in AI output
  • Provide context that counters potential biases (e.g., specifying gender)
  • Report problematic outputs
The Copyright Question: Where Did the Training Data Come From?

AI models were trained on vast amounts of text scraped from the internet, including copyrighted books, articles, and creative works. This has sparked a wave of lawsuits.

The Legal Landscape (2026)

What We KnowDetails
Lawsuits filed70+ copyright infringement cases
Mixed rulingsSome courts say "fair use," others disagree
Biggest settlement$1.5 billion (2025)
Legal statusStill uncertain

Why This Matters to You

For commercial use: The legal status of AI-generated content remains uncertain. Safest approach: treat AI output as a starting point that you substantially modify.

The broader debate: Some creators feel their work was used without consent or compensation. There are no easy answers, but awareness helps you make informed choices.

Privacy: Where Does Your Conversation Go?

When you chat with AI, what happens to your data? It varies by provider and changes over time. Always read the current privacy policy.

General Patterns

Your conversations may be stored:

  • Most providers keep logs (at least temporarily)
  • Used for service improvement and safety monitoring

Your data might train future models:

  • Some providers use conversations for training
  • Others offer opt-out options
  • Review policies for your specific provider

Sensitive information carries risk:

  • Even with strong policies, data could be accessed through breaches or legal processes

Practical Guidelines

Never share with AI:

  • Passwords
  • API keys
  • Credentials

Be cautious with:

  • Confidential business information
  • Personal health details
  • Anything you would not want to become public

Action item: Check your provider's privacy settings and opt-out options.

Environmental Cost: AI Uses a Lot of Energy

Training and running AI requires substantial energy. Here are the numbers:

The Big Picture

MetricValue
US electricity to data centers4.4%
AI data center CO2 (past year)105 million metric tons
Share of national emissions2%+
Comparison**Exceeds aviation industry**

Per-Query Impact

  • Single AI query: ~0.3 watt-hours
  • Sounds tiny, but billions of queries daily adds up
  • 60% of increased demand met by fossil fuels

Context Matters

  • Training is far more energy-intensive than querying
  • Once trained, running AI is relatively efficient
  • Streaming video, crypto mining consume comparable energy
  • AI companies increasingly investing in renewables

What You Can Do

  • Use AI for tasks with genuine value (not trivial queries)
  • Support providers investing in sustainability
  • Recognize "AI is bad for the environment" oversimplifies a complex tradeoff
Misinformation and Deepfakes: AI-Generated Deception

AI can generate convincing text, images, audio, and video. This enables misinformation at unprecedented scale.

Types of AI-Powered Misinformation

Text misinformation:

  • Fake news articles, social media posts, comments
  • Difficult to distinguish from human-written content
  • Bad actors can produce misinformation faster and cheaper

Visual misinformation:

  • Photorealistic fake images
  • Fake video footage of real people
  • Convincing voice cloning

The authentication problem:

  • Proving content is authentic becomes harder
  • Erodes trust broadly, even in genuine content

What You Can Do

  • Be skeptical of content designed to provoke strong emotions
  • Look for original sources and corroboration
  • Support journalism and fact-checking organizations
  • Do not share content you have not verified
  • Remember: "Seeing is believing" no longer applies
Your Responsibility as an AI User

Using AI has implications beyond your immediate task. Principles for thoughtful use:

1. Verify before sharing You are responsible for content you put your name on, regardless of how it was created.

2. Disclose when appropriate Norms are evolving, but deceiving others about AI involvement is problematic. Academic settings, professional contexts, and creative submissions often have explicit policies.

3. Consider impact on others If AI helps you work faster, are you making appropriate use of those gains? Could your AI use affect others' livelihoods?

4. Stay informed AI capabilities, policies, and ethical considerations evolve rapidly.

Bottom line: Use AI like any powerful tool. Thoughtfully, with awareness of both benefits and risks.

Now that you have a solid foundation, you are ready to actually use it. Next unit: your first real AI conversation.

Example Prompts to Try

Prompt 1: Privacy Inquiry

What happens to this conversation after we finish? Is it used to train you?

Note: AI might not have current info about its own provider's policies. Verify with official documentation.

Prompt 2: Bias Test

Generate a professional bio for a software engineer named Alex.

Look for gender assumptions. Does AI use he/him, she/her, or they/them? What other assumptions appear?

Prompt 3: Ethical Complexity

What are the ethical concerns I should consider when using AI? Be specific and balanced.

Tests whether AI can engage thoughtfully with complexity vs. defensive dismissal or overwrought caution.

Hands-On Exercise

Goal: Research your AI tool's actual privacy policy.

Step 1: Go to the official website (claude.ai, openai.com, etc.)

Step 2: Find their privacy policy or data usage documentation.

Step 3: Find answers to these questions:

  • Are conversations stored? For how long?
  • Is data used to train AI models?
  • Can you opt out of data training?
  • Can you delete conversation history?
  • What data is collected beyond messages?

Step 4: Compare what you found to what you expected. Anything surprising?

This ensures you make informed choices based on actual policies, not assumptions.

Key Takeaways
  • AI inherits biases from training data. Be aware and question assumptions, especially on sensitive topics.
  • Your conversations may be stored and used. Read privacy policies and avoid sharing sensitive information.
  • AI has real environmental costs. Use it for genuine value, not trivial queries.
  • AI enables sophisticated misinformation. Be skeptical, verify sources, and do not trust content just because it looks professional.
  • You are responsible for how you use AI. Verify before sharing, disclose when appropriate, and stay informed.
Sources