COURSE 29L0100% FREE
Verified 2026-08-10

The dangers of AI — why we study the bad with the good

### What this Course is

What this Course is

This Course is a light, honest map of harms and misuse already linked to AI in the real world—plus risks experts publicly debate.

It is not a tutorial. You will not find recipes, attack steps, exploit code, or detailed visuals of abuse. The goal is civic literacy: if society only markets the shiny demos, people cannot make informed choices.

Why it matters

AI can draft emails, catch fraud, translate languages, and help scientists. The same families of tools have also been used—or feared—for scams, impersonation, biased decisions, harassment, and more. Understanding both sides reduces blind fear and blind trust.

Public agencies already treat some AI-enabled harms as current problems, not science fiction. For example, the U.S. Federal Trade Commission has highlighted AI-driven impersonation and deepfake-related fraud risk in consumer protection work (FTC, Feb 2024 ↗).

How to read these chapters

Each chapter answers roughly:

  1. What is the danger category? (plain words)
  2. What has already shown up in public reporting or research? (high level)
  3. Why should an ordinary person care?
  4. What healthy responses look like? (skepticism, reporting, policy literacy—not DIY crime)

If you want defensive technical depth (secure design, evaluation, fairness metrics), use Course 18 and Course 19. Those Courses teach protection and measurement, not attacks.

What we refuse to publish here

  • Step-by-step instructions for causing harm
  • Copy-paste jailbreak or exploit recipes aimed at producing dangerous content
  • Detailed imagery of violence, exploitation, or intimate deepfakes
  • Any code samples on Course 29 pages

Try it

Before reading further chapters, write two lists: AI hopes (3 items) and AI worries (3 items). After finishing Course 29, revisit the worry list: which items are documented today, which are speculative, which you still cannot judge?

Myths

⚠️ Myth: Talking about dangers teaches people to commit crimes.
✓ Reality: Public health and safety education routinely name risks without providing crime manuals. We follow that pattern.
⚠️ Myth: If a risk is listed, it means every AI tool does it.
✓ Reality: Misuse depends on tools, incentives, safeguards, and human choices.
⚠️ Myth: Only future “superintelligence” matters.
✓ Reality: Present-day fraud, bias, and overtrust already harm people.

Sources