AI Safety Engineering
Implement guardrails, red-team prompts, and build safer AI applications.
0%
Confidence · no data yet
After this skill you can…
- Implement input and output guardrails
- Red-team a deployed LLM application
- Use Llama Guard or NeMo Guardrails
Prerequisites
Watch (10 videos)
Prompt Injection Explained: How AI Agents Get Tricked!
→ Implement safety controls→ Test AI systems for security risks
Auditable AI Tools: Scalable Governance for Next-Gen AI Systems
→ Design safe AI systems→ Conduct AI audits→ Develop explainable AI models
Anthropic Sounds AI Alarm
→ Design safety protocols for AI systems→ Develop strategies for human-AI collaboration→ Analyze potential risks of autonomous AI training
Are we creating new patient safety risks in the name of opioid reduction?
→ Design individualized multimodal analgesia care→ Mitigate drug-drug interactions
Safety Test Behind Every Parachute
→ Design safety protocols→ Implement safety testing→ Analyze material weaknesses
Don’t Roll Out XR Training Until You Have This in Place
→ Design Safe XR Training Programs→ Mitigate Risks in XR Training
Why Frontier AI Models Still Show Bias — And Why It’s Harder to Fix Than People Think
→ Design and implement AI systems that minimize bias→ Develop strategies for mitigating bias in AI models→ Evaluate the safety and fairness of AI systems
Why Cancer Detection AI Keeps Failing Doctors in 2026
→ Design and implement safety protocols for AI systems→ Evaluate AI systems for potential risks and biases
Asbestos is a bigger problem than we thought
→ Design safety measures for asbestos handling→ Implement protocols for mitigating asbestos risks→ Develop emergency response plans for asbestos exposure
Group Acupuncture in Cancer Care: NHS Experience with Mandy Brass, MSc MBAcC
→ Design safe healthcare systems
DeepCamp AI