{ "platformId": "chatgpt-work", "taskSlug": "task-18868", "directAnswer": "Use ChatGPT Work to build a consistent qualification-to-requirement evidence matrix from the same authorized job criteria and application fields for every candidate. It may cite where evidence supports, contradicts, or does not establish a requirement and flag criteria that may act as unlawful or irrelevant proxies. Keep the matrix from becoming a verdict. Missing evidence is not a failed qualification; polished prose is not stronger evidence; and similarity to prior hires is not job relatedness. ChatGPT must not infer protected traits, rank people, reject applicants, advance a shortlist, contact candidates, or write back to the applicant system. The attached figures are projected role-task priors, not measured ChatGPT adoption by human resources specialists, so they support a controlled trial rather than a claim about current practice.", "topUseCases": [ { "interventionId": "task-18868:screening-evidence-matrix:recommend", "taskId": "task-18868", "task": "Review employment applications and job orders to match applicants with job requirements.", "workflowId": "screening-evidence-matrix", "authorityMode": "recommend", "centralityScreen": { "value": 0.12638960879999997, "importance": { "value": 0.7949999999999999, "sourceField": "role-tasks-kv:task-18868.importance" }, "occupationalRelevance": { "value": 0.9308, "sourceField": "role-tasks-kv:task-18868.relevance / 100" }, "frequency": { "kind": "modal-response-share", "modalResponseSharePct": 17.08, "collectionDate": "2026-06-03", "sourceField": "role-tasks-kv:task-18868.modalResponseSharePct" } }, "agentFit": { "value": 0.9025, "sourceField": "rule:digital-feasibility-and-capability-fit-single-count.v2" }, "distribution": { "value": 0.9, "sourceField": "roles-kv:aiPlaybook.delegatedTasks[task-18868].projected-role-task-prior" }, "evidenceConfidence": "weak", "withinGateScore": 0.10265995974779997, "gate": { "status": "limited", "reason": "Authority remains below an approval-required intervention." }, "capabilityIds": [ "connected-apps-files", "deliverable-generation" ], "expectedArtifact": "Qualification-to-requirement evidence matrix, same format and audit trail for every candidate", "humanReviewPoint": "The shortlist and every selection decision are yours with documented rationale; the matrix is evidence, not a verdict.", "riskLevel": "high" } ], "workflows": [ { "id": "screening-evidence-matrix", "goal": "Turn applications and job requirements into an evidence matrix that supports a human selection decision without becoming one.", "inputs": [ "Job requirements", "Anonymized applications", "Your selection-decision protocol" ], "connectedSystems": [ "Applicant tracker export", "Document repository" ], "steps": [ "Match qualifications to requirements with cited evidence per candidate, same format for all.", "Work from qualifications only; flag any requirement criterion that could proxy a protected characteristic.", "Produce the audit trail: what was compared, what evidence supported each match, what was not evaluated.", "Stop at the matrix: ranking candidates for rejection is declined; the decision protocol is human." ], "expectedArtifact": "Per-candidate evidence matrix with audit trail and proxy-criterion flags", "starterPrompt": "Here are the job requirements and anonymized applications. Match qualifications to requirements with cited evidence per candidate, do not infer or report protected characteristics; flag job criteria that may operate as proxies, and give me the audit trail in the same format for everyone. The shortlist decision is mine; if I ask you to rank and reject the bottom half, decline and give me the evidence matrix instead.", "humanReviewPoint": "This guide treats selection as a human decision with documented rationale; adverse-impact hygiene is structural in the matrix, not optional.", "commonFailure": "A matrix that quietly becomes a ranking, or requirement criteria that proxy protected traits passing through unflagged." } ], "setupRecommendations": [ { "item": "Use a fixed, approved requirement rubric; provide only minimum-necessary, preferably anonymized application evidence through read-only sources; keep ranking, status-change, and candidate-contact permissions disabled.", "rationale": "Applicant matching is defensible only when job relatedness, evidence rules, and treatment are fixed before review and every candidate receives the same comparison without protected-trait inference.", "capabilityIds": [ "connected-apps-files", "approval-enterprise-controls" ] } ], "prompts": [ { "id": "screening-guardrails", "title": "Screen with guardrails", "prompt": "Here are the job requirements and anonymized applications. Match qualifications to requirements with cited evidence per candidate; do not infer or report protected characteristics; flag job criteria that may operate as proxies; and give me the audit trail. The shortlist decision is mine.", "expectedArtifact": "Evidence matrix with audit trail and proxy-criterion flags", "humanReviewPoint": "You make the shortlist with documented rationale." } ], "humanOwnedResponsibilities": [ "Job-relatedness law: humans approve each requirement, its necessity, evidence rule, weighting, and any lawful accommodation before screening", "Same-treatment law: the same rubric, source fields, evidence states, and review procedure apply to every candidate in the comparison population", "No-inference law: protected characteristics, health, family status, age, origin, religion, disability, and proxy traits are neither inferred nor used", "Missing-evidence law: absent or ambiguous application evidence is recorded as NOT ESTABLISHED, not converted into a negative fact or automatic rejection", "Selection law: shortlist, interview, rejection, accommodation, adverse-action, communication, and applicant-system updates remain human decisions and actions" ], "risks": [ { "id": "task-18868-scaffold-risk", "level": "high", "description": "An evidence matrix can disguise inconsistent treatment, convert missing information into failure, import proxy criteria, or become an automated ranking that determines who receives an opportunity.", "mitigation": "Freeze and approve a job-related rubric before review, minimize and anonymize inputs, use identical evidence states for all candidates, audit proxy criteria and population treatment, and require documented human selection decisions." } ], "sources": [ { "id": "occupation-human-resources-specialists", "title": "Human Resources Specialists occupational render model", "url": "https://singulariki.com/roles/human-resources-specialists", "publisher": "gold.onet_role_starter_surface", "lastReviewedAt": "2026-06-03" }, { "id": "openai-chatgpt-work", "title": "ChatGPT Work and Codex", "url": "https://help.openai.com/en/articles/20001275-chatgpt-work-and-codex", "publisher": "OpenAI", "lastReviewedAt": "2026-07-11" }, { "id": "openai-work-deliverables", "title": "Creating and editing documents, spreadsheets, and presentations with ChatGPT Work", "url": "https://help.openai.com/en/articles/20001278-creating-and-editing-documents-spreadsheets-and-presentations-with-chatgpt-work", "publisher": "OpenAI", "lastReviewedAt": "2026-07-11" }, { "id": "openai-work-cloud-browser", "title": "Using cloud browser in ChatGPT", "url": "https://help.openai.com/en/articles/20001280-using-cloud-browser-in-chatgpt", "publisher": "OpenAI", "lastReviewedAt": "2026-07-11" }, { "id": "openai-work-desktop-browser", "title": "Using the built-in browser in the ChatGPT desktop app", "url": "https://help.openai.com/en/articles/20001277-using-the-built-in-browser-in-the-chatgpt-desktop-app", "publisher": "OpenAI", "lastReviewedAt": "2026-07-11" }, { "id": "openai-chatgpt-skills", "title": "Skills in ChatGPT", "url": "https://help.openai.com/en/articles/20001066", "publisher": "OpenAI", "lastReviewedAt": "2026-07-11" }, { "id": "openai-chatgpt-scheduled-tasks", "title": "Scheduled Tasks in ChatGPT", "url": "https://help.openai.com/en/articles/10291617-tasks-in-chatgpt", "publisher": "OpenAI", "lastReviewedAt": "2026-07-11" }, { "id": "openai-workspace-agents", "title": "ChatGPT Workspace Agents for Enterprise and Business", "url": "https://help.openai.com/en/articles/20001143", "publisher": "OpenAI", "lastReviewedAt": "2026-07-11" } ], "freshness": { "platformVersion": "2026-07-09-launch", "platformReviewedAt": "2026-07-11", "occupationalDataVersion": "O*NET 30.3", "occupationalDataGeneratedAt": "2026-06-03T05:05:52.989984+00:00" }, "quality": { "uniqueOccupationalOrTaskEvidence": false, "validPlatformSpecificWorkflow": true, "clearOutputArtifact": true, "realCapabilityMatch": true, "meaningfulHumanReviewSection": true, "sourcesAndFreshness": false, "distinctInternalRelationships": true, "noUnresolvedFactualOrProductVersionFailure": false, "indexable": false } }