Tiếp theoHướng dẫn tiếp theo
Luật địa phương 144 của NYC và Kiểm tra thiên vị tuyển dụng AI
xã hội
HƯỚNG DẪN KỸ THUẬT
An AI bias audit evaluates whether a system’s decisions or errors differ across relevant groups and whether the differences are justified for the use.
Nghe hướng dẫn nàyAI thuật lại · ~3 min
A strong audit defines the decision and affected population, selects metrics connected to the harm, tests subgroups and intersections, documents limitations, and leads to corrective action. No single metric can establish fairness in every context.
Begin with the decision, system boundary, and people affected. Identify what the model predicts, how a score changes a real decision, the deployment setting, and which group differences could cause harm. A bias audit is not just a single fairness metric: it should document data sources, selection stages, target definitions, model versions, decision thresholds, and human review. Choose metrics tied to the harm, such as selection rates, false-positive or false-negative rates, calibration, or error severity. Different metrics can conflict, so explain why the chosen measures fit the decision. Disaggregate results by legally and contextually relevant groups, including intersections where sample size and privacy permit. Report counts and uncertainty, not only percentages. Examine data coverage, proxy features, missingness, and whether labels measure the intended construct. The 2019 Obermeyer study, for example, found a health-management algorithm that used cost as a proxy for need assigned lower risk scores to Black patients with comparable illness. An audit should test the target and process, not just the model’s final output. For employment tools covered by New York City Local Law 144, employers and employment agencies must arrange an independent bias audit within one year before use, publish a summary, and provide required notices. That specific local law does not define every AI audit everywhere. Other laws and policies can require different methods or records. An auditor should disclose scope, data limitations, metrics, and whether the system was tested under conditions matching actual use. Close the loop: identify findings, assign remediation owners, decide whether to alter data, model, threshold, or process, and retest before relying on the system. Keep a dated report and version history. A passing metric is not proof of fairness, legal compliance, or validity. If critical gaps remain, limit use or do not deploy until evidence and safeguards are adequate.
A New York City employer checks whether its covered AEDT has a recent independent bias audit and publishes the required summary before use.
A health system tests whether a care-management score uses past spending as a proxy for need, informed by the 2019 Obermeyer study.
A lender evaluates approval and error patterns using lawfully obtained or carefully estimated demographic data and documents limitations of any proxy method.
A face-verification provider reports error rates for intersectional groups instead of relying only on overall accuracy.
Repeat the review after changes in the model, data, threshold, vendor, workflow, or jurisdiction. Pair scheduled testing with a named owner for complaints, monitoring signals, and remediation deadlines. Check local rules separately because audit independence, publication, and notice requirements vary. When fixes change the model or decision policy, measure results against the original harm and test for new group disparities before expanding use. Preserve prior reports so teams can compare outcomes over time. Preserve reviewer approvals alongside reports. Keep change logs.
Bản tin miễn phí
Ba câu chuyện AI đã được xác minh vào mỗi buổi sáng các ngày trong tuần, được viết bằng tiếng Anh đơn giản. Miễn phí mãi mãi, không có quảng cáo.
Một email mỗi ngày trong tuần. Hủy đăng ký trong một cú nhấp chuột. Chúng tôi không bao giờ bán hoặc chia sẻ địa chỉ của bạn.
Tự kiểm tra
Phản hồi tức thì về mọi câu trả lời và chứng chỉ có thể chia sẻ với ID có thể xác minh sau khi bạn vượt qua khóa học.
Hỗ trợ giáo dục AI miễn phí. AI Understanding là tổ chức phi lợi nhuận 501(c)(3) — không bao giờ có quảng cáo, không có tường phí. Hãy quyên góp
An AI bias audit evaluates whether a system’s decisions or errors differ across relevant groups and whether the differences are justified for the use. A strong audit defines the decision and affected population, selects metrics connected to the harm, tests subgroups and intersections, documents limitations, and leads to corrective action. No single metric can establish fairness in every context.
The guide recommends defining the decision and affected population before selecting metrics.
Counts and uncertainty help reviewers interpret noisy estimates for smaller groups.
The study found that using cost as a proxy for need understated need for Black patients with comparable illness.
NYC DCWP says covered tools require a recent bias audit, public summary, and notices.
A passing metric does not establish fairness, validity, or legal compliance in every respect.
Tiếp tục học hỏi
Đã chọn thêm hướng dẫn cho chủ đề này
Tiếp theoHướng dẫn tiếp theo
Luật địa phương 144 của NYC và Kiểm tra thiên vị tuyển dụng AI
xã hội