Governance and accountability
Governance evidence
What agencies reported about the OMB M-25-21 minimum practices for high-impact AI. This is a disclosure analysis, not an audit, a compliance finding or an official assessment. An unanswered question is missing evidence, not evidence that a safeguard is absent.
Practice-by-practice evidence
Denominator: 7 records · deployed high-impact use casesReported in placeReported in progressNot applicablePrecluded by law or guidanceOther or ambiguous answerNo answer reported
Pre-deployment testing
Pre-deployment testing
Reported in place: 4 (57.1%)Pre-deployment testing
Reported in progress: 2 (28.6%)Pre-deployment testing
No answer reported: 1 (14.3%)
7Reported in place: 4 (57.1%)Pre-deployment testing
Reported in progress: 2 (28.6%)Pre-deployment testing
No answer reported: 1 (14.3%)
AI impact assessment
AI impact assessment
Reported in place: 3 (42.9%)AI impact assessment
Reported in progress: 3 (42.9%)AI impact assessment
No answer reported: 1 (14.3%)
7Reported in place: 3 (42.9%)AI impact assessment
Reported in progress: 3 (42.9%)AI impact assessment
No answer reported: 1 (14.3%)
Independent review
Independent review
Reported in place: 3 (42.9%)Independent review
Reported in progress: 3 (42.9%)Independent review
No answer reported: 1 (14.3%)
7Reported in place: 3 (42.9%)Independent review
Reported in progress: 3 (42.9%)Independent review
No answer reported: 1 (14.3%)
Ongoing monitoring
Ongoing monitoring
Reported in place: 4 (57.1%)Ongoing monitoring
Reported in progress: 2 (28.6%)Ongoing monitoring
No answer reported: 1 (14.3%)
7Reported in place: 4 (57.1%)Ongoing monitoring
Reported in progress: 2 (28.6%)Ongoing monitoring
No answer reported: 1 (14.3%)
Operator training
Operator training
Reported in place: 3 (42.9%)Operator training
Reported in progress: 3 (42.9%)Operator training
No answer reported: 1 (14.3%)
7Reported in place: 3 (42.9%)Operator training
Reported in progress: 3 (42.9%)Operator training
No answer reported: 1 (14.3%)
Fail-safe
Fail-safe
Reported in place: 3 (42.9%)Fail-safe
Reported in progress: 2 (28.6%)Fail-safe
Not applicable: 1 (14.3%)Fail-safe
No answer reported: 1 (14.3%)
7Reported in place: 3 (42.9%)Fail-safe
Reported in progress: 2 (28.6%)Fail-safe
Not applicable: 1 (14.3%)Fail-safe
No answer reported: 1 (14.3%)
| Practice | In scope | In place | In progress | Not in place | N/A | Precluded or waived | Other | No answer | Answered | In place of applicable |
|---|---|---|---|---|---|---|---|---|---|---|
| Pre-deployment testing | 7 | 4 | 2 | 0 | 0 | 0 | 0 | 1 | 86% | 57% of 7 |
| AI impact assessment | 7 | 3 | 3 | 0 | 0 | 0 | 0 | 1 | 86% | 43% of 7 |
| Independent review | 7 | 3 | 3 | 0 | 0 | 0 | 0 | 1 | 86% | 43% of 7 |
| Ongoing monitoring | 7 | 4 | 2 | 0 | 0 | 0 | 0 | 1 | 86% | 57% of 7 |
| Operator training | 7 | 3 | 3 | 0 | 0 | 0 | 0 | 1 | 86% | 43% of 7 |
| Fail-safe | 7 | 3 | 2 | 0 | 1 | 0 | 0 | 1 | 86% | 50% of 6 |
| Appeal process | 7 | 2 | 2 | 0 | 1 | 1 | 0 | 1 | 86% | 40% of 5 |
| User and public consultation | 7 | 2 | 3 | 0 | 0 | 0 | 1 | 1 | 86% | 29% of 7 |
Answered = records with any answer ÷ records in scope. In place of applicable = reported in place ÷ (in scope minus not applicable, precluded and waived). Unanswered records stay in that denominator, so the rate is a lower bound on reported evidence, not an estimate of real practice.
| Testing | Impact assess. | Indep. review | Monitoring | Training | Fail-safe | Appeal | Consultation | n | |
|---|---|---|---|---|---|---|---|---|---|
| DHS | 100% | 67% | 67% | 100% | 67% | 100% | 100% | 67% | 3 |
| DOJ | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 2 |
| VA | – | – | – | – | – | – | – | – | 1 |
| SSA | 100% | 100% | 100% | 100% | 100% | 100% | 100% | 0% | 1 |
Hatched cells: the agency answered nothing for that practice, so there is no evidence either way. Compare agencies with care: each agency is one reporting unit, and portfolios differ in size and type.
Records requiring further investigation
4 records with at least one practice unanswered or in progressOpen in explorer| Use case | Stage | Test | AIA | Review | Monitor | Train | Fail-safe | Appeal | Consult |
|---|---|---|---|---|---|---|---|---|---|
| Verification Match Model DHS · DHS-2384 | Deployed | Reported in place | Reported in progress | Reported in progress | Reported in place | Reported in progress | Reported in place | Reported in place | Reported in progress |
| Medical Claims Adjudication DOJ · DOJ-0156 | Deployed | Reported in progress | Reported in progress | Reported in progress | Reported in progress | Reported in progress | Reported in progress | Reported in progress | Reported in progress |
| Truview DOJ · DOJ-0168 | Deployed | Reported in progress | Reported in progress | Reported in progress | Reported in progress | Reported in progress | Reported in progress | Reported in progress | Reported in progress |
| Automated Decision Support VA · VA-24-4722 | Deployed | No answer reported | No answer reported | No answer reported | No answer reported | No answer reported | No answer reported | No answer reported | No answer reported |
Unanswered analytical questions
Generated from the data; for follow-up with agencies- VA designated 3 use cases high-impact but answered none of the minimum-practice questions for any of them. Were the practices assessed, and where is that recorded?
- 3 high-impact use cases are reported as deployed while their AI impact assessment is reported in progress. Was deployment authorised before the assessment was complete?
- 1 high-impact use cases mark an appeal process as not applicable. On what basis was each individual outcome judged not contestable?
- 3 high-impact use cases report using PII but give no link to a Privacy Impact Assessment. Is a PIA published elsewhere?
These are questions the inventory raises but cannot answer. They are not findings of non-compliance.