MOC

AI Safety & Ethics

Updated 2 May 2025
mocsafetyethicsalignmentsecurityai

AI Safety & Ethics

AI safety isn’t one thing. It’s three things that people often conflate:

  1. Practical security — The threats happening right now. Deepfakes, scams, prompt injection. Real people, real harm, today.
  2. Alignment & bias — Making AI systems do what we actually want, fairly. The alignment problem, systemic bias, technical approaches to safety.
  3. Existential risk — The long-term question of whether advanced AI poses a fundamental threat to humanity. Serious people disagree about this.

All three matter. This section covers all three.

Practical AI Security

The threats that exist right now and what to do about them.

  • AI Security — Start here. The threat landscape: what’s real, what’s coming, how to defend
  • Deepfakes — Fake images, video, and audio of real people. Detection, prevention, legal status
  • AI Scams & Social Engineering — Voice cloning scams, AI-powered phishing, fake identities. The new economics of fraud
  • Prompt Injection — The #1 security vulnerability in AI applications. How it works, why it’s hard to fix

This section connects directly to regulation and case law. The harm is real and the law is responding.

Alignment & Bias

The harder, slower problems. How do we make AI systems that are fair, honest, and aligned with human values?

Existential Risk

The biggest questions. Not everyone agrees these risks are real. We present the arguments.

Learning Paths

Structured journeys through this material, tailored to who you are:

  • AI Safety Courses — Guided learning paths:
    • For everyone — protect yourself (30 min)
    • For businesses — govern and comply (2 hrs)
    • For developers — build securely (2 hrs)
    • The philosophy path (1.5 hrs)

Why This Section Exists

Because the people building AI and the people affected by AI need different things from the same conversation.

If you’re deploying AI: You need to know about Prompt Injection, the EU AI Act, and bias. These are operational risks with legal consequences.

If you’re a citizen: You need to know about Deepfakes, scams, and how to protect yourself and your family. This is personal safety.

If you’re thinking about the future: You need to grapple with AI Alignment and existential risk. These are the questions that will define the next century.

If you’re a regulator: You need to understand all of the above — the technology, the threats, the philosophical debates — to make good law. See Legal & Compliance and Regulator Watch.

Go Deeper

enes