Skip to content
ClelandCo

AI risk

Updated August 15, 2026 · 3 min read

How to red-team an AI system before it ships

A practical rehearsal guide for prompt injection, poisoned data, dependency failure, bias, misuse, and recovery, with evidence and scope limits.

By , M.S. Artificial IntelligencePublished Updated

The short answer

Red-teaming an AI system means deliberately attacking, degrading, and misusing it under controlled conditions before it carries material work. Six useful exercise families are prompt injection, data poisoning, failure injection, bias evaluation, misuse simulation, and a disaster-recovery walkthrough. They are not exhaustive; the threat model determines which cases matter and what evidence each test must produce.

Start with what you are testing for

Red-teaming without a threat model and decision path can produce observations with no agreed triage or remediation action. Before running anything, write down what this specific system could do that would actually hurt: what data it can reach, what actions it can take without a human, whose decisions it influences, and what the worst plausible consequence of a confident wrong answer is. The exercises below are then aimed at those, rather than run as a generic checklist.

A red-team report is only useful if every finding comes with a reproduction. A list of concerns is not a finding.

The six exercises

1. Prompt injection

Place hostile instructions where the system may read them as ordinary input: inside a retrieved document, web page, support ticket, filename, or calendar invite. Test both direct instructions and indirect instructions embedded in content the workflow was designed to trust; the relevant question is what the whole system does with the manipulated output.

What you are measuring is not really whether the model can be talked out of its instructions. It is whether anything downstream acts on the result without a check. A model that can be manipulated into producing bad text is a nuisance; a model that can be manipulated into producing bad text that then triggers an API call is an incident.

2. Data poisoning

Introduce corrupted or adversarial records into a retrieval corpus, fine-tuning set, or feedback loop. Include both intentional manipulation and ordinary operational failures such as a bad import, mislabeled batch, or feedback path that amplifies incorrect data.

Measure whether bad inputs are detectable before they are trusted, how far a poisoned record propagates, which outputs depend on it, and whether the record can be removed and the affected artifacts rebuilt.

3. Failure injection

Remove or degrade a dependency. Make the API return errors, latency, a truncated response, and valid-looking nonsense. The last case deserves its own assertion because it can bypass outage handling while still degrading the decision.

The question is whether the system degrades safely or fails open, and — separately — whether the documented fallback path has ever actually been executed. A fallback that exists only in a design document is not a fallback.

4. Bias and disparate impact

Evaluate against a held-out set that reflects the populations the system is expected to encounter, including underrepresented groups where the data supports responsible analysis. Report performance by relevant group as well as in aggregate, because an aggregate can conceal materially different error patterns.

Then make the legal, ethical, and operational judgment explicitly: where performance differs, by how much, what uncertainty surrounds the estimate, and whether the difference has consequences in this specific use. The same measured gap can carry very different risk in different decisions.

5. Misuse simulation

Use the system the way busy people will actually use it. Take the shortcuts. Paste in the customer data. Use it for the adjacent task it was not designed for but is obviously convenient for. The point is not to catch anyone out — it is to find out what the sanctioned path costs in friction.

If the unsanctioned path is materially easier, the incentive to route around policy rises. The exercise identifies that friction so the approved path and the control can be redesigned together.

6. Disaster-recovery walkthrough

Run the full scenario as a tabletop: the model was wrong, someone acted on it, a customer noticed before you did. Walk it end to end with the people who would actually be involved.

Measure time to detect and contain, who is called, what is disclosed and to whom, and whether the logs needed to reconstruct the model output and downstream action still exist.

What to do with the findings

Sort them into three buckets: fix before launch, accept with a documented rationale, and monitor with a named threshold. The middle bucket is legitimate and under-used — some risks are worth carrying, and writing down why is what makes it a decision rather than an oversight. What is not legitimate is a finding that lands in no bucket, which is how a red-team report becomes a liability with a date on it.

References and boundaries

Primary references, not borrowed authority.

These sources inform the risk and governance framing. They do not endorse ClelandCo, validate a client outcome, or turn this practitioner guide into a certification standard.

Questions

Asked and answered.

How long does this take?
It depends on the threat model, number of integrations, actions the system can take, data sensitivity, and number of cases that must be reproduced and retested. A scoped plan should name the exercise set, participants, evidence format, remediation owner, and retest window rather than promise a generic duration.
Can our own team run this?
Yes. The builders know the architecture and likely failure paths, but they can share assumptions that an independent reviewer will challenge. Pair internal knowledge with someone who did not design the system, and make every finding reproducible so disagreement can be resolved against evidence.
Is this the same as a penetration test?
No. A pen test covers network and application security and is a different discipline with different certifications. This covers the model, the data around it, and the workflow it sits in. Where a finding is a conventional security issue, it should be handed to people who do that for a living.

Design the rehearsal around the actual system.

Start with the threat model, safe test environment, evidence record, stop conditions, and authorized people—not a generic attack checklist.