A New Floor for Entry-Level Talent
The entry-level tech job market has bifurcated. On one side: traditional programming roles that have contracted sharply — overall programmer employment in the U.S. fell 27.5% between 2023 and 2025, according to labour statistics tracked by IEEE Spectrum. On the other side: information security analyst and AI engineering roles, which have seen double-digit growth in the same window.
What is causing this split is not a general contraction of tech hiring. Tech employment overall is projected to reach 9.8 million in the U.S. by 2026, representing 1.9% growth. The contraction is specific to roles where the task profile has been substantially automated — code completion, boilerplate writing, routine test generation — and the expansion is in roles that require directing, evaluating, and extending AI systems.
Universities are responding to this split structurally. Purdue University announced in early 2026 that AI working competency — the ability to select, evaluate, and use AI tools appropriately within a discipline — would become a requirement for graduation, starting with incoming cohorts in fall 2026. Purdue is among the most prominent institutions to formalise this requirement, but dozens of other universities have introduced mandatory AI literacy modules, updated core curriculum requirements, or embedded AI tool proficiency assessments into capstone projects. The practical effect is that graduates entering the market in 2026 and 2027 will have a documented AI baseline that previous cohorts lacked.
This has direct implications for how hiring managers should approach entry-level hiring. The class of 2026 is not the same as the class of 2022 — not only because it was educated in a more AI-rich environment, but because it was formally assessed on AI competency as a condition of graduation.
What the Data Actually Shows About Entry-Level AI Skills
The picture from hiring data is more nuanced than either “AI is replacing all entry-level jobs” or “nothing has changed.” NACE (National Association of Colleges and Employers) data shows that 61% of employers say they are not replacing entry-level jobs with AI, while 41% are discussing or planning to augment those jobs with AI within the next five years. The dominant pattern is augmentation, not replacement — but augmentation requires entry-level employees who can work with AI tools effectively, which raises the minimum bar for every junior role.
According to CompTIA’s 2026 workforce research, 275,000+ active U.S. job postings referenced AI skills in January 2026. AI/ML engineering roles surged 163% year-over-year. Entry-level hiring at the 15 biggest U.S. tech firms dropped 25% from 2023 to 2024 — but this reflects a shift toward fewer, higher-qualified entry-level hires rather than an exit from entry-level hiring. Employers’ rating of the 2026 job market for new graduates is at its most pessimistic since 2020, though 49% still view it as “good” or “very good.”
The divergence is between two profiles: graduates who can demonstrate applied AI proficiency — building pipelines, evaluating model outputs, using AI tools to accelerate their existing specialisation — and graduates who cannot. The first profile is increasingly well-compensated and in demand. The second profile is competing for a reduced pool of traditional roles against a larger field of similarly qualified candidates.
Advertisement
What Hiring Managers Should Do About It
The university-driven baseline shift creates both a challenge and an opportunity for hiring managers. The challenge: traditional job descriptions and screening criteria may now be filtering out the candidates most worth hiring. The opportunity: the new AI-competent graduate cohort represents a different value proposition than entry-level hires have historically offered.
1. Rewrite Entry-Level Job Descriptions to Match the New Graduate Profile
Job descriptions that do not mention AI tools, prompt engineering, or the ability to evaluate model outputs are sending the wrong signal. Graduates who have been assessed on AI competency as a graduation requirement will deprioritise roles that appear to be from the pre-AI era. More practically, ATS systems filtering for “AI” or “LLM” in submitted resumes will not surface these candidates for roles that do not include those terms — creating a self-fulfilling skills gap where companies cannot find AI-literate candidates for roles that did not ask for AI literacy.
The rewrite does not require a full role transformation. Adding a line — “Familiarity with AI coding assistants and ability to critically evaluate AI-generated outputs” — signals current relevance and attracts the cohort that was trained against that standard. For technical roles, specifying which AI tools or frameworks are in use signals honesty about the actual working environment.
2. Build Assessment Methods That Measure Applied AI Competency, Not Just Knowledge
Many companies have not updated their interview and assessment processes for the new baseline. A coding interview that prohibits AI tools is assessing a skill set that does not match how the job will actually be done. A take-home project that requires building a feature in 48 hours without AI assistance is testing historical performance, not forward performance. The companies that will attract the best AI-competent graduates are the ones whose hiring processes signal that they operate in a realistic AI-augmented environment — using tools like GitHub Copilot, Claude, or domain-specific AI assistants — and that they want candidates who can work effectively within that environment.
This is practically manageable: a pair-programming interview session that includes AI tools, or a take-home project where candidates are explicitly permitted and encouraged to use AI assistance, produces more relevant signal about on-the-job performance than an AI-prohibited equivalent.
3. Recalibrate Expectations for What Entry-Level Contribution Looks Like
The AI-augmented junior developer is not the same as the 2020-era junior developer. A junior developer with effective AI coding assistance can produce code volumes and architectural decisions that previously required 2-3 years of experience. This creates a tricky calibration problem: compensation expectations have not fully adjusted to reflect this productivity boost, but the contribution potential has. Companies that capture the most value from AI-competent junior talent will treat them as higher-leverage contributors from day one, give them harder problems earlier, and invest in the evaluation skills (code review of AI-generated output, system design judgment) that complement AI assistance rather than duplicate it.
The Antitrust Question in Credentials
The formalisation of AI competency in graduation requirements has a second-order effect that hiring managers should track: credential inflation. When AI literacy becomes a baseline requirement for graduation, the credential loses differentiation value quickly. The class of 2025 that voluntarily added AI skills to their profile had a genuine competitive advantage. The class of 2027 where AI competency is mandatory will not — because everyone in the pool will have met the same baseline.
This mirrors what happened to general programming skills post-2010: once “knows Python” went from differentiator to expectation, the differentiation moved to what you built with Python. In 2028 and beyond, the differentiation will move from “can use AI tools” to “built something specific and demonstrably useful with AI tools.” Hiring managers who start evaluating portfolios of AI-assisted projects now — rather than waiting for credentials to catch up — will have a more accurate signal about candidate quality before the credential baseline fully compresses.
Frequently Asked Questions
What are universities actually requiring when they mandate “AI competency” for graduation?
AI competency requirements vary by institution and discipline. At Purdue and comparable institutions, the baseline includes: the ability to select an appropriate AI tool for a given task; the ability to critically evaluate AI-generated outputs for accuracy, bias, and appropriateness; and the ability to use AI assistance ethically, with appropriate attribution. For technical disciplines, this typically extends to proficiency with AI coding assistants and the ability to review and debug AI-generated code. The requirements are assessed through capstone projects and integrated assignments, not standalone AI certifications.
How does the rise of AI competency requirements affect graduate pay expectations?
The short-term effect is upward pressure on entry-level salaries for AI-competent graduates, as supply of genuinely AI-proficient candidates remains below demand. AI/ML engineering roles command $134,000–$193,300 annually at the top of the market, and even junior roles in AI-augmented environments are being compensated at higher levels than equivalent non-AI roles. The medium-term effect (2028-2030) is that AI competency becomes table stakes, and the premium compresses as the baseline rises. Companies that invest in retention of AI-competent junior talent now — before the baseline fully rises — will benefit from lower churn and higher contribution margins.
Should hiring managers at smaller companies with limited AI experience still adjust their entry-level criteria?
Yes — and the adjustment is simpler than it sounds. For companies that are not themselves AI-native, the key change is removing barriers that discourage AI-competent candidates: prohibitions on AI tool use in coding interviews, job descriptions that make no mention of AI tools, and onboarding processes that do not include AI tooling. The graduate who was assessed on AI competency as a graduation condition will self-select away from environments that appear to operate in the pre-AI era. Small adjustments in how roles are described and how interviews are conducted produce disproportionate improvements in the calibre of AI-competent applicants who engage.
—
