AI Safety Courses
AI Safety Courses
Learning paths through the hub’s safety and security content. Not traditional courses with modules and quizzes — guided journeys through the material, sequenced so each page builds on the last.
Pick the path that matches who you are and what you need.
For Everyone — Protect Yourself
Time: ~30 minutes of reading Who it’s for: Anyone who uses the internet, has family members, or exists in 2025.
This isn’t about technical AI knowledge. It’s about recognising the new landscape of AI-powered threats and knowing how to protect yourself and the people you care about.
| Step | Page | What You’ll Learn |
|---|---|---|
| 1 | AI Security | Overview of the threat landscape — what’s real, what’s hype |
| 2 | AI Scams & Social Engineering | How AI-powered scams work and how to defend against them |
| 3 | Deepfakes | How fake media is created, spotted, and used for harm |
| 4 | Audio & Speech AI | Understand voice cloning — what makes it dangerous |
After this path you’ll know:
- How to spot AI-generated content
- How to protect yourself and family from voice cloning scams
- When to be sceptical of media you see online
- Practical steps you can take today
For Businesses — Govern and Comply
Time: ~2 hours of reading Who it’s for: Business leaders, compliance officers, risk managers, anyone deploying or procuring AI.
The question isn’t whether to use AI — it’s how to use it safely, legally, and responsibly.
| Step | Page | What You’ll Learn |
|---|---|---|
| 1 | AI Security | The threat landscape for your organisation |
| 2 | AI Scams & Social Engineering | Threats targeting your employees |
| 3 | Prompt Injection | The #1 technical vulnerability if you’re building with AI |
| 4 | EU AI Act | Your legal obligations (if you serve EU users) |
| 5 | GDPR & AI | Data protection meets AI — the tensions and requirements |
| 6 | ISO 42001 AI Management System | The certifiable AI governance framework |
| 7 | AI Bias & Fairness | Reputational and legal risk from biased AI systems |
| 8 | AI Alignment | Why you should care about the alignment problem |
After this path you’ll know:
- What AI risks your organisation faces
- What the law requires of you
- How to assess AI vendors
- How to build an AI governance programme
- When you need human oversight
For Developers — Build Securely
Time: ~2 hours of reading Who it’s for: Software developers building AI-powered applications, integrations, or agents.
You write the code. You’re the last line of defence.
| Step | Page | What You’ll Learn |
|---|---|---|
| 1 | Prompt Injection | The vulnerability you MUST understand |
| 2 | AI Security | The full landscape of what can go wrong |
| 3 | AI Agents | How autonomy multiplies security risk |
| 4 | Agent Frameworks | Secure patterns for agent development |
| 5 | AI Alignment | Why your design choices matter for alignment |
| 6 | AI Bias & Fairness | How bias enters your system and how to mitigate it |
| 7 | EU AI Act | What the law expects of AI developers |
| 8 | How LLMs Work | Understanding the system you’re defending |
After this path you’ll know:
- How to defend against prompt injection (and why there’s no silver bullet)
- Secure patterns for AI application development
- Where to add human-in-the-loop checkpoints
- How to think about AI safety as a development discipline
- Your legal obligations as a developer
The Philosophy Path
Time: ~1.5 hours of reading Who it’s for: Anyone interested in the deeper questions about AI safety — alignment, existential risk, the nature of the problem.
| Step | Page | What You’ll Learn |
|---|---|---|
| 1 | AI Alignment | The core challenge: making AI do what we actually want |
| 2 | Constitutional AI | One approach to alignment (Anthropic’s) |
| 3 | AI Bias & Fairness | The concrete, present-day manifestation of alignment failure |
| 4 | Existential Risk from AI | The long-term stakes |
| 5 | Geoffrey Hinton | The “Godfather of AI” who now warns about what he built |
| 6 | Ilya Sutskever | The researcher who left OpenAI over safety concerns |
How These Paths Connect
Every path overlaps. The individual learning about scams benefits from understanding prompt injection. The developer benefits from the business governance perspective. The businessperson benefits from understanding the technical foundations.
That’s by design. This hub is a graph, not a book. These courses are suggested routes through the graph, not walled gardens.
Go Deeper
- AI Safety & Ethics — The section home
- AI Security — Start here for practical security
- Legal & Compliance — The regulatory landscape
- Regulator Watch — What regulators are doing right now
- AI Intelligence Hub — Back to the hub home