Technology landscape
How reported AI classifications spread across agencies, missions, lifecycle stages and sourcing. Each record carries one reported classification, so categories do not overlap. A classification describes the technique, not a specific commercial model.
Lifecycle by technology
Share of each technology's recordsDeployed: 69 (69.0%)Generative AI
Pilot: 23 (23.0%)Generative AI
Pre-deployment: 7 (7.0%)Generative AI
Retired: 1 (1.0%)
Deployed: 8 (80.0%)Classical / predictive ML
Pilot: 1 (10.0%)Classical / predictive ML
Pre-deployment: 1 (10.0%)
Sourcing by technology
Where reportedDeveloped in-house: 9 (9.0%)Generative AI
Contract and in-house: 28 (28.0%)Generative AI
Purchased from vendor: 63 (63.0%)
Developed in-house: 1 (10.0%)Classical / predictive ML
Contract and in-house: 3 (30.0%)Classical / predictive ML
Purchased from vendor: 6 (60.0%)
Contract and in-house: 6 (42.9%)Natural language processing
Purchased from vendor: 8 (57.1%)
Technology adoption by agency
Row share: each agency's records by classification · 16 largest portfolios in the filter| Generative AI | Classical / predictive ML | Natural language processing | Computer vision | Agentic AI | Reinforcement learning | Other | n | |
|---|---|---|---|---|---|---|---|---|
| DOE | 85% | 2% | 7% | 2% | 2% | 2% | 2% | 59 |
| DOJ | 46% | 31% | 15% | 8% | 13 | |||
| HHS | 75% | 25% | 8 | |||||
| SSA | 86% | 14% | 7 | |||||
| DHS | 83% | 17% | 6 | |||||
| STATE | 50% | 33% | 17% | 6 | ||||
| FTC | 100% | 5 | ||||||
| DOL | 100% | 4 | ||||||
| DOI | 25% | 50% | 25% | 4 | ||||
| SBA | 100% | 4 | ||||||
| USDA | 67% | 33% | 3 | |||||
| EPA | 67% | 33% | 3 | |||||
| HUD | 50% | 50% | 2 | |||||
| EAC | 100% | 2 | ||||||
| NRC | 100% | 1 | ||||||
| FHFA | 100% | 1 |
Technology by topic area
Row share within each topic| Generative AI | Classical / predictive ML | Natural language processing | Computer vision | Agentic AI | Reinforcement learning | Other | n | |
|---|---|---|---|---|---|---|---|---|
| Information Technology | 86% | 5% | 5% | 2% | 2% | 56 | ||
| Administrative Functions | 64% | 7% | 7% | 14% | 4% | 4% | 28 | |
| Other | 75% | 8% | 17% | 12 | ||||
| Energy and the Environment | 100% | 7 | ||||||
| Science | 43% | 29% | 29% | 7 | ||||
| Government Benefits Processing | 80% | 20% | 5 | |||||
| Law Enforcement | 60% | 40% | 5 | |||||
| Human Resources | 33% | 67% | 3 | |||||
| Service Delivery | 67% | 33% | 3 | |||||
| Cybersecurity | 100% | 2 | ||||||
| Procurement and Financial Management | 100% | 2 | ||||||
| Emergency Management | 100% | 1 | ||||||
| Health and Medical | 100% | 1 | ||||||
| Transportation | 100% | 1 |
High-impact designation by technology
High-impact: 4 (4.0%)Generative AI
Presumed high-impact, determined not: 2 (2.0%)Generative AI
Not high-impact: 93 (93.0%)Generative AI
Not reported: 1 (1.0%)
High-impact: 1 (7.1%)Natural language processing
Not high-impact: 13 (92.9%)
Governance disclosure by technology
High-impact records: reported in place / scope| Technology | High-impact | Testing | Impact assess. | Indep. review | Monitoring | Training |
|---|---|---|---|---|---|---|
| Generative AI | 4 | 2 / 4 | 4 / 4 | 4 / 4 | 1 / 4 | 4 / 4 |
| Classical / predictive ML | 0 | – | – | – | – | – |
| Natural language processing | 1 | 0 / 1 | 0 / 1 | 0 / 1 | 0 / 1 | 0 / 1 |
| Computer vision | 0 | – | – | – | – | – |
| Agentic AI | 0 | – | – | – | – | – |
| Reinforcement learning | 0 | – | – | – | – | – |
| Other | 0 | – | – | – | – | – |
Shows counts reported in place against all high-impact records of that technology. The remainder includes in-progress, not applicable and unanswered. Full governance view
Systems, vendors and technology are different things
133 of 133 records name a vendor (100%)Vendor: the supplier the agency names in a free-text field. Matched to a standard list of companies; one record may name several.
System name: the agency's name for the system (reported for 87% of records).
AI classification: the technique category the agency selected. A vendor name does not show which model or technique is used, and this platform does not infer it.