The federal government is hiring for artificial intelligence. The question is whether the current approach will get the results agencies are hoping for.
The Office of Personnel Management (OPM) authorized government-wide direct-hire authority for AI positions in December 2023, effective through 2028. This covers several occupational series: IT specialist (2210), computer scientist — artificial intelligence (1550), computer engineer — artificial intelligence (0854), and management and program analyst (0343). The authority covers grades GS-9 through GS-15.
OPM followed up with a memo outlining pay and benefits flexibilities. Agencies can offer recruitment incentives up to 25% of basic pay times the number of years in a service agreement, capped at four years. Relocation incentives for existing employees follow the same formula. Retention incentives are available at up to 25% of basic pay for a single employee or 10% for a group. OPM noted that for these incentives, the existence of direct-hire authority itself serves as justification that a position is hard to fill.
The Department of Homeland Security (DHS) launched a "hiring sprint" to build a team of 50 AI experts for its AI Corps, using the direct-hire authority to streamline the process.
Meanwhile, agencies are investing heavily in upskilling their existing workforce. FedScoop reported that 95% of agencies are focused on training current employees to work with AI. This tracks with GitLab research showing 84% of public sector DevSecOps professionals expect AI to significantly affect their work in the next five years.
But Federal News Network raised a sharp question in a June 2026 interview. Dr. Priyanka Dave, the Upskilling Lead in the Administrative Modernization Program at Oregon State University, suggested that agencies may be solving the wrong problem. The issue is not a lack of training courses. It is a lack of organizational support for applying new skills.
"Employees need to feel safe when they are practicing new things or when they are learning new things," Dave said. The core problem is cultural. If managers penalize mistakes during the learning curve, employees won't experiment with AI tools. If agencies measure productivity during the experimentation phase, they will train people without seeing any performance improvement.
The upskilling critique lands at a critical moment. The federal government is spending significant resources on AI talent. But new skills need new systems — not just new courses.
The practical takeaway: if you are a GovCon leader building an AI competency, focus on the environment around the training. Create safe spaces for experimentation. Measure learning outcomes, not just course completions. And keep an eye on the direct-hire competitions — they are pulling AI talent in-house faster than many contractors expect.
Accuracy Review — Post 5
- Confirm OPM direct-hire authority date: December 2023 (not December 2026 as some sources imply vs. FedScoop reporting)
- Verify incentive percentages: up to 25% of basic pay for recruitment/relocation, 4-year cap — confirm against OPM guidance via FedScoop
- Confirm DHS AI Corps "hiring sprint" for 50 positions — verify source and date
- Verify "95% of agencies focused on upskilling" claim — FedScoop references this; confirm the exact study/source
- Verify Dr. Priyanka Dave's quote and affiliation on Federal News Network — note the full transcript was not retrieved; check that "solving the wrong problem" framing is attributed correctly
- Confirm GS-9 through GS-15 grade range and occupational series (2210, 1550, 0854, 0343) against OPM direct-hire memo
- Note that some search results returned conflicting direct-hire dates (2023 vs. 2024) — verify exact OPM memo date
Posts written in TCB brand voice (Flesch-Kincaid grade 7-9, sentences 15-20 words, 2-3 sentence paragraphs, no banned AI-isms). Each post closes with an actionable takeaway for federal IT and GovCon professionals.