Role Guides

AI Risk by Role

Different users face different AI risks. These role guides translate the same failure modes for students, developers, cybersecurity analysts, business leaders, and compliance or risk teams.

Showing 5 role guides

Choose your role

Students Developers Cybersecurity Analysts Business Leaders Compliance and Risk Teams

Students

Students often use AI for studying, writing, brainstorming, summarizing, and research. The biggest risk is treating fluent AI output as verified knowledge.

Top risks

HallucinationFabricated CitationsOutdated KnowledgeFalse PrecisionOverconfidence

Recommended checklists

Before You Trust an AI Answer
Before You Paste Sensitive Data

Practical advice

Use AI to explain concepts, generate study questions, and outline ideas, but verify important claims independently.
Never submit AI-generated citations unless you personally checked that the sources exist and support the claim.
Ask the AI to explain uncertainty, assumptions, and what should be verified.
Use AI as a tutor or brainstorming partner, not as a replacement for learning the material.

Developers

Developers face risks from plausible but broken code, insecure patterns, outdated APIs, dependency issues, and tool-connected automation mistakes.

Top risks

Non-runnable CodeVersion MismatchInsecure CodeTool MisuseSensitive Data ExposurePrompt Injection

Recommended checklists

Before You Use AI for Code
Before You Paste Sensitive Data

Practical advice

Run AI-generated code locally before trusting it.
Check imports, package versions, file paths, and framework assumptions.
Do not paste secrets, internal source code, credentials, or production logs into unapproved AI tools.
Treat AI code suggestions like an external pull request that requires review.

Cybersecurity Analysts

Cybersecurity analysts can use AI to accelerate triage and documentation, but AI can invent CVEs, overstate threats, miss context, or expose sensitive incident data.

Top risks

HallucinationFabricated CitationsPrompt InjectionSensitive Data ExposureOverconfidenceTool Misuse

Recommended checklists

Before You Use AI for Cybersecurity Research
Before You Paste Sensitive Data
Before You Trust an AI Answer

Practical advice

Verify CVEs, CWEs, KEV status, vendor advisories, and exploit claims against authoritative sources.
Separate confirmed exploitation from theoretical attack paths.
Avoid sharing sensitive logs, internal architecture, tickets, or incident details with unapproved AI systems.
Use AI to assist analysis, not to replace analyst judgment or source verification.

Business Leaders

Business leaders need to understand where AI can improve productivity while also creating risks from overtrust, privacy leakage, vendor dependence, and weak review processes.

Top risks

OverconfidenceAutomation BiasSensitive Data ExposureTool MisuseVendor Lock-inOutdated Knowledge

Recommended checklists

Before You Trust an AI Answer
Before Your Team Deploys an AI Tool
Before You Paste Sensitive Data

Practical advice

Define where AI may assist and where human review is required.
Set clear rules for sensitive data, customer information, and confidential business material.
Start with low-risk pilots before expanding AI use across teams.
Require accountability for decisions influenced by AI output.

Compliance/Risk Teams

Compliance and Risk Teams

Compliance and risk teams need practical controls for AI use, including data handling rules, review processes, documentation, vendor evaluation, and governance boundaries.

Top risks

Sensitive Data ExposurePrompt InjectionTool MisuseOverconfidenceCitation LaunderingModel Update Breakage

Recommended checklists

Before Your Team Deploys an AI Tool
Before You Paste Sensitive Data
Before You Trust an AI Answer

Practical advice

Document approved and prohibited AI use cases.
Create data classification rules for AI prompts and outputs.
Require review and approval for high-impact decisions influenced by AI.
Track vendor claims, data retention policies, security controls, and model-update risks.