---
title: "Assemble the record for an automated-decision impact assessment"
url: "https://ailawframework.com/obligations/ai-decision-systems/draft-a-privacy-impact-or-algorithmic-impact-assessment-for-the-system/united-states"
type: "task_jurisdiction_version"
site: "AI Law Framework"
id: "A45-A20260923-2-US"
practice_area: "AI decision systems (the firm's and clients')"
tier: "yellow"
lawyer_review: "required"
jurisdictions: ["US"]
last_reviewed: "2026-10-22"
---

# Assemble the record for an automated-decision impact assessment

> From the system documents the lawyer identifies, list the stated purpose, the data categories, the decisions the output is used for, and any human-review step the documents describe.…

Canonical page: https://ailawframework.com/obligations/ai-decision-systems/draft-a-privacy-impact-or-algorithmic-impact-assessment-for-the-system/united-states  
Last reviewed: 2026-10-22

Site index for agents: https://ailawframework.com/llms.txt

## Key facts

| Fact | Value |
| --- | --- |
| Practice area | [AI decision systems (the firm's and clients')](https://ailawframework.com/obligations/ai-decision-systems.md) |
| Human involvement | Yellow: Lawyer and AI draft together |
| How AI is used | From the system documents the lawyer identifies, list the stated purpose, the data categories, the decisions the output is used for, and any human-review step the documents describe. Do not write the assessment's conclusions. |
| What the lawyer must review | If the system is used to recruit, screen, hire, monitor, set pay, promote or fire, the EEOC's 29 April 2024 page says the federal employment-discrimination statutes apply to that use, including a neutral tool with an unjustified disparate impact. Colorado Senate Bill 26-189 was signed on 14 May 2026. The bill page's summary says that, starting 1 January 2027, a developer of covered automated decision-making technology that materially influences a consequential decision must give the deployer technical documentation, and that deployers must give consumers notice and keep compliance records for at least three years. Those dates are not yet in force on 22 October 2026. You read the signed act before treating a duty as applicable, and you write the assessment. Do not put personal data into a public tool. Formal Opinion 512 requires verification and, for a self-learning tool, the Rule 1.6 consent analysis. |
| Written for | United States |
| Applies to | [Draft a privacy-impact or algorithmic-impact assessment for the system](https://ailawframework.com/obligations/ai-decision-systems/draft-a-privacy-impact-or-algorithmic-impact-assessment-for-the-system.md) |

## Checking plan

- What the AI does: From the system documents the lawyer identifies, list the stated purpose, the data categories, the decisions the output is used for, and any human-review step the documents describe. Do not write the assessment's conclusions.
- Basis: Checked against the linked sources on 22 October 2026. Confirm the current instrument before reliance.

## Notes

- The Colorado summary defines a consequential decision as one about education, employment, housing, financial or lending services, insurance, health-care services, or essential government services and public benefits. Confirm that text in the signed PDF.
- There is no single federal impact-assessment statute cited here. Do not invent one.

## Sources

- [EEOC, What is the EEOC's role in AI? (29 April 2024)](https://www.eeoc.gov/sites/default/files/2024-04/20240429_What%20is%20the%20EEOCs%20role%20in%20AI.pdf) (checked 2026-10-22)
- [Colorado Senate Bill 26-189, signed act 14 May 2026](https://leg.colorado.gov/bills/sb26-189) (checked 2026-10-22)
- [ABA Formal Opinion 512, Generative Artificial Intelligence Tools (29 July 2024)](https://www.americanbar.org/content/dam/aba/administrative/professional_responsibility/ethics-opinions/aba-formal-opinion-512.pdf) (checked 2026-10-22)

## Forms and precedents for this task

| Form or precedent | Kind | AI level | Summary |
| --- | --- | --- | --- |
| [Law firm generative AI use policy (United States)](https://ailawframework.com/precedents/ai-decision-systems/us-law-firm-generative-ai-use-policy.md) | Policy | Yellow | An internal policy for a US law firm on which generative AI tools may be used, what may be put into them, who must review output, and when the client is told.… |
| [Approval and training record for a generative AI tool (United States)](https://ailawframework.com/precedents/ai-decision-systems/us-generative-ai-tool-approval-record.md) | Checklist | Yellow | The record a US firm completes before a generative AI tool is put on the permitted list, covering confidentiality settings, retention, reliability, liability, notice of breach, and who has been… |
| [Record that a person reviewed an automated output](https://ailawframework.com/precedents/ai-decision-systems/human-review-of-automated-output.md) | Form | Yellow | Short record that a named person reviewed one automated output before anyone acted on it, what they checked, and whether they departed from it.… |
| [Reassessment when a model, purpose or provider changes](https://ailawframework.com/precedents/ai-decision-systems/model-or-purpose-change-reassessment.md) | Checklist | Yellow | Checklist for when the model, the purpose, the data or the provider of a decision system changes, asking which earlier conclusions have to be made again and by whom. |

Signed-in accounts can read the full text of each one.

*Practice guidance for planning and review. Confirm current primary sources and the requirements of the actual matter. This is general information, not legal advice.*