Risk-based approach
A risk-based approach prioritizes decisions, requirements, controls, or resources according to assessed risks and potential consequences.
Introduction to Generative AI, Second Edition
A risk-based approach is a regulatory method that assigns requirements according to the potential severity and likelihood of harm from a system.
- How it is used
- Regulates AI applications by matching oversight and obligations to their potential societal harm.
- How deep it goes
- Defines risk categories and details high-risk controls, while leaving lower-risk requirements less fully specified.
- Examples
- Uses a violent voice-activated children’s toy and applications in hiring, education, law enforcement, and financial services.
- What you can do after
- Evaluate AI applications by risk level and identify corresponding bans, assessments, documentation, oversight, and governance duties.
Read how Introduction to Generative AI, Second Edition explains risk-based approach →
Data-Driven Cybersecurity
A decision-making approach that prioritizes cybersecurity actions and resources according to assessed risk and potential impact.
- How it is used
- Aligns cybersecurity efforts and investments with business goals and executive priorities.
- How deep it goes
- Defines risk-based prioritization and explains executive reporting, business risk assessments, and investment decisions without detailing a formal scoring method.
- Examples
- Contrasts counting blocked phishing attempts with communicating attack impact and security-training return on investment.
- What you can do after
- Evaluate security investments by translating threats into financial and operational risks for executive decisions.
Read how Data-Driven Cybersecurity explains risk-based approach →
The Art of AI Product Development
Risk-aware decision making is the practice of evaluating potential consequences and applying stricter controls to high-stakes actions.
- How it is used
- Guides AI agents to escalate high-stakes actions before execution rather than discovering consequences after failure.
- How deep it goes
- Defines consequence-sensitive controls and gives a three-step finance example, without specifying implementation details.
- Examples
- Uses ambiguous fund transfers, transaction thresholds, previous-transaction checks, and multistep verification.
- What you can do after
- Operate agents with approval gates, anomaly checks, verification steps, rollback mechanisms, and audit logs.
Read how The Art of AI Product Development explains risk-based approach →
Related concepts
- AI Actused by · The AI Act is a European Union regulation establishing harmonised, risk-based requirements for the development, placing on the market, and use of artificial intelligence systems.
- guardrailuses · A guardrail is a constraint, safeguard, or control that limits actions or behavior to reduce risk and keep a system or process within acceptable bounds.
- riskuses · Risk is the possibility and potential impact of an adverse outcome arising from uncertainty.
- risk assessmentuses · Risk assessment is the systematic identification and evaluation of potential adverse events, including their likelihood and consequences, to determine their significance.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.