Businesses are adopting AI faster than they are securing it. AI systems can be attacked through the data they learn from, the models themselves and the applications built around them. This course teaches your technical team how to build and defend AI systems against attacks.
Who it is for
Security practitioners, data scientists, machine learning engineers and developers who build, deploy or defend AI systems. Attendees need experience or education beyond junior level.
Format
1 to 5 days, on-site or online, for up to 10 people, built from the modules you choose.
Modules
- How AI systems work, for security practitioners
- Adversarial inputs and evasion attacks
- Model theft, extraction and privacy attacks
- Supply chain risks in pre-trained models, datasets and third-party AI services
- Prompt injection, jailbreaks and other attacks on language models
- Securing AI agents and the tools they can use
- The lethal trifecta of private data, untrusted content and external communication in AI agents
- AI security frameworks, including MITRE ATLAS and the OWASP Top 10 for LLM applications
- How attackers use AI, and what it changes for defenders
- Testing your own AI systems with AI red teaming
What you will be able to do
- Assess an AI system, model or agent for security weaknesses
- Build security controls into AI systems, from data collection to deployment
- Test your own AI systems before an attacker does
Pricing
- Priced per team
- From £1,750 a day for up to 4 people
- From £2,750 a day for up to 10 people1
- Prices exclude VAT
- See pricing for every option
-
Discounts: 15% off for businesses based on the Isle of Man and 10% off for retainer clients, up to 25% combined. Micro businesses with 4 or fewer employees get heavily discounted pricing, so contact us for details. See the terms. ↩︎