AWS Certified AI PractitionerResponsible AIHard
A novel AI system is being developed to predict the likelihood of recidivism (reoffending) for individuals in the justice system. The system's predictions could influence sentencing, parole decisions, and resource allocation. Given the profound impact on individuals' lives and potential for societal harm, a multidisciplinary team is conducting an extensive evaluation covering ethical, legal, and social implications, including potential biases, fairness concerns, privacy risks, and accountability mechanisms, before any deployment. Which crucial Responsible AI practice is this team undertaking?
- AData anonymization
- BEthical Impact Assessment (EIA)
- CAI model fine-tuning
- DAdversarial training
Show answer & explanationAnswer & explanation
Correct answer: B. Ethical Impact Assessment (EIA)
The extensive, multidisciplinary evaluation of ethical, legal, and social implications, including fairness, privacy, and accountability, prior to deployment for a high-impact AI system, is the definition and purpose of an Ethical Impact Assessment (EIA).
Why the other options are wrong
- A. Data anonymization is a privacy technique, not a holistic impact assessment.
- C. AI model fine-tuning is a technical process to improve performance, not an ethical evaluation.
- D. Adversarial training is a technique for improving model robustness, not an ethical assessment.
Ethical Impact Assessment (EIA)
An Ethical Impact Assessment (EIA) is a systematic process to identify, analyze, and mitigate the ethical, legal, and societal risks and opportunities associated with the development and deployment of an AI system.
- Conducted before deployment, especially for high-stakes AI.
- Covers fairness, privacy, accountability, transparency, and societal impact.
- Involves multidisciplinary teams and stakeholder consultation.
Memory trick: Ethical AI checks before it leaps.