Curriculum / Module 08
professional2 hours3 lessonsTimed assessment

The Adversarial Frontier

Poisoning, prompt injection, manipulated inputs

This module deals with security. Earlier modules dealt with models that are misused, or that fail on their own; the concern here is models that are deliberately attacked by someone who wants a particular result. It covers three attack classes: adversarial evasion, training-data poisoning, and prompt injection against AI-integrated applications. Each is connected to the financial threat of manipulated inputs submitted to fraud and underwriting models. The module returns to Module 5 to show how a manipulated application can defeat an automated credit decision, and it sets out a realistic defensive posture: what can be hardened, what cannot, and where a person must remain in the decision path.

Create an accountBack to curriculum

By the end of this module, you can
  • Explain evasion, poisoning and prompt-injection attacks in plain terms, and identify the point in a financial workflow at which each of them takes effect.
  • Recognise manipulated or adversarial inputs aimed at fraud, credit and monitoring models.
  • Describe a realistic defensive posture, state what defence cannot achieve, and identify the decisions on which a person must remain in the decision path.
Skills taught
  • Adversarial-threat literacy
  • Input-integrity checking
  • Recognising indirect prompt injection in AI-assisted tools
  • Setting realistic expectations of model robustness

Lessons in this module

L1
Evasion: adversarial examples and gamed inputs
Explain adversarial evasion in plain terms, connect it to manipulated inputs submitted to fraud, credit and monitoring models, and explain why complete robustness cannot be achieved.
30 min
L2
Poisoning: corrupting what the model learns
Explain training-data and feedback poisoning, identify where financial models are exposed to it, and connect it to the label-quality lessons of Modules 4 and 6.
25 min
L3
Prompt injection: attacking AI-integrated applications
Explain direct and indirect prompt injection, identify where AI-integrated financial tools are exposed, and specify the containment posture, since injection cannot be fully solved by prompting.
30 min
Case 08.1
Case file: The application that learned the model
An applied fact pattern worked against a model resolution, followed by the timed assessment (30 minutes, pass mark 70 percent).
25 min

How it lands across the four desks

Fraud & AML

The models you rely on are themselves targets: evasion attacks tuned to pass below a detection threshold, poisoning that corrupts what the model learns from its own case outcomes, and manipulated inputs built to score clean. You learn to recognise the patterns each attack leaves in ordinary monitoring data, and what your defences can and cannot do about them.

Credit & Underwriting

This module returns to the question Module 5 left open: how a manipulated application, built from adversarially tuned inputs and gamed features, defeats an automated credit decision, and why checking the integrity of the inputs submitted with an application is now part of underwriting.

Compliance & Risk

You govern the controls and set the institution's expectations of model robustness: what a vendor can honestly promise, what the phrase 'adversarially robust' does and does not mean, and the circumstances in which keeping a person in the decision path is a security control rather than a matter of preference.

Wealth & Advisory

The AI assistants you adopt form part of the attack surface as well, because indirect prompt injection can be carried in the documents and web content you ask a tool to process. You learn where that risk sits before you delegate a task to a tool that reads material from outside the institution.

Key literature · 7 sources

Every module rests on a verified scholarly and institutional evidence base. The full core and further reading lists open with the module.

  • Szegedy, C. et al. (2014) 'Intriguing Properties of Neural Networks.' ICLR 2014: the discovery of adversarial examples.
  • Goodfellow, I. J., Shlens, J. & Szegedy, C. (2015) 'Explaining and Harnessing Adversarial Examples.' ICLR 2015: the mechanism and the reliability/transferability of evasion.
  • Biggio, B. & Roli, F. (2018) 'Wild Patterns: Ten Years After the Rise of Adversarial Machine Learning.' Pattern Recognition 84: the field's consolidated, sober consensus.
  • NIST (2025) 'Adversarial Machine Learning: A Taxonomy and Terminology' (AI 100-2e2025): the shared vocabulary for evasion, poisoning and their variants.
  • Greshake, K. et al. (2023) 'Not What You've Signed Up For: Compromising Real-World LLM-Integrated Applications with Indirect Prompt Injection.' AISec '23: the naming and demonstration of indirect injection.
  • OWASP GenAI Security Project (2025) 'OWASP Top 10 for LLM Applications': prompt injection and related integration risks as leading concerns.
Create an account