The AI Jobs Reckoning: Why a Tech CEO Says One in Three New Graduates Could Be Unemployed

On March 13, 2026, ServiceNow CEO Bill McDermott appeared on CNBC's "Squawk on the Street" and delivered a warning that crystallized the anxieties of an entire generation: unemployment for new college graduates "could easily go into the mid-30s in the next couple of years" as AI agents devour the entry-level workforce [1]. The prediction, coming from the head of a $130 billion enterprise software company that has itself automated 90% of its customer service operations, was not the speculation of a bystander. It was a confession from the front lines.

McDermott's remarks arrive at a moment when the abstract fear of AI-driven job displacement is hardening into concrete data, legislative alarm, and a wave of layoffs that has made the first quarter of 2026 a turning point in the conversation about artificial intelligence and work.

The Numbers Behind the Warning

The current unemployment rate for recent college graduates already tells a troubling story. According to the Federal Reserve Bank of New York, unemployment among recent graduates climbed to approximately 5.7% in the fourth quarter of 2025, while the underemployment rate—measuring graduates working in jobs that typically do not require a college degree—surged to 42.5%, the highest level since the pandemic year of 2020 [4][5]. In the first quarter of 2025, graduate unemployment hit 5.8%, the highest reading since 2021 [3].

These figures represent a sharp reversal. For most of the post-pandemic recovery, the labor market favored graduates. Now, since the national unemployment rate fell back below 5%, new graduates have continuously faced a higher unemployment rate than the overall U.S. rate [4]. As of June 2025, new graduates faced 4.8% unemployment compared to the national rate of 4.0% [4].

Source: Bureau of Labor Statistics / Federal Reserve Bank of New York

Data as of Mar 13, 2026CSV

The overall U.S. unemployment rate has drifted upward from 3.4% in April 2023 to 4.4% in February 2026, according to Bureau of Labor Statistics data [6]. But for recent graduates, the deterioration has been far steeper. The gap between college graduate unemployment and high school graduate unemployment narrowed to just 2.5 percentage points in mid-2025—the smallest advantage since records began tracking the comparison [4]. The college degree, once a reliable economic shield, is losing its protective power.

"So Much of the Work Is Going to Be Done by Agents"

McDermott's specific concern centers on agentic AI—autonomous software systems that can execute multi-step workflows without human intervention. "So much of the work is going to be done by agents," he told CNBC. "It's going to be challenging for young people to differentiate themselves in the corporate environment" [1].

This is not hypothetical. ServiceNow, which sells workflow automation software to major enterprises, reported that 97% of standard software at the company is now automatically generated by AI, and 80% of customer inquiries are fully managed by agentic AI [2]. The company's AI capabilities generated $600 million in annual contract value in 2025, with a target of $1 billion for 2026 [7]. Each of ServiceNow's 28,000 employees has undergone an AI skill assessment [2].

McDermott's dual role—warning about displacement while selling the tools that enable it—underscores the paradox at the heart of this crisis. ServiceNow's tools "will help businesses slash hiring costs," he told CNBC in the same interview [1]. The company's stock surged 16% in the days surrounding the interview as investors embraced its "Autonomous Roaming" strategy [7].

A Wave of AI-Driven Layoffs

The warning lands amid the most significant wave of AI-linked corporate restructuring since the technology emerged. Days before McDermott's interview, Atlassian announced it was cutting 1,600 jobs—roughly 10% of its workforce—explicitly to "self-fund" investments in AI and enterprise sales [8]. That followed Block's announcement that it would eliminate approximately 4,000 of its 10,000 positions, making it the largest AI-driven layoff of 2026 [8][9].

These are not isolated actions. Across industries, companies are pivoting toward workflows where AI systems handle a majority of operations and customer support tasks that once required large entry-level teams [9]. The automation is falling hardest on repetitive cognitive roles: manual testing, technical support, basic coding, data entry, and routine software maintenance [10].

A Stanford Digital Economy Lab study analyzing millions of payroll records found that employment for software developers aged 22 to 25 declined nearly 20% from its late-2022 peak, while employment for those over 30 held steady or grew [11]. Entry-level hiring at the top 15 technology firms fell 25% from 2023 to 2024 alone [11]. Research showed that after late 2022, AI-adopting companies hired five fewer junior workers per quarter than before [10].

Source: GDELT Project

Data as of Mar 13, 2026CSV

The Chorus of Alarm

McDermott is far from alone in sounding the alarm. In November 2025, U.S. Senator Mark Warner (D-Virginia) warned that unemployment for recent college graduates could surge to 25% within two to three years, triggering a "level of social disruption that's unprecedented" [12]. Warner's biggest fear, he said, was not the layoffs visible in headlines but "the jobs that will never be created," noting that "many big banks have shared with me they are cutting interns and first-year hires in half" [12].

Warner joined with Senator Josh Hawley (R-Missouri) to introduce bipartisan legislation requiring agencies and major companies to deliver quarterly reports to the Department of Labor on AI's impact on their workforces [13]. A companion bill, the AI Workforce PREPARE Act, would establish a research hub within the Department of Labor to evaluate AI's workforce effects, while authorizing up to $90 million in grants to support displaced workers and $160 million for technology education programs [13].

Anthropic CEO Dario Amodei has forecast that 50% of entry-level jobs may be wiped out as AI technology improves [10]. In February 2026, Fortune reported on the emerging consensus that even a 10% reduction in employment "will feel like a depression" given the social and psychological centrality of work to American identity [14].

The Talent Doom Cycle

Not everyone agrees that replacing junior workers with AI is wise corporate strategy. AWS CEO Matt Garman called the practice "one of the dumbest things I've ever heard," warning that companies should continue hiring graduates and questioning how organizations will function "10 years in the future when no one has built up or learned anything" [11].

CNBC reported on what workforce researchers call a "talent doom cycle": when companies eliminate entry-level positions, they destroy the pipeline that produces future mid-level and senior talent [11]. Healthy organizations cultivate their own leadership internally. If you remove too many junior roles, you starve the internal talent pipeline, forcing expensive external hires and creating dependency on an increasingly competitive talent market [11].

The concern runs deeper than organizational logistics. Entry-level jobs have historically served as training grounds where graduates develop judgment, institutional knowledge, and the tacit skills that cannot be taught in a classroom. When AI handles debugging, testing, basic coding, and data analysis, junior employees "interact with outputs rather than problems," compressing experience but weakening the foundational skill development that produces competent senior professionals [11][15].

Microsoft executives have echoed this concern, warning that AI could "limit the developer talent pipeline" if companies fail to maintain entry-level hiring [11].

A Generation Under Pressure

The data on who is being hit hardest is stark. An analysis of job postings reveals that 35% of positions labeled "entry-level" now require years of prior relevant work experience—a phenomenon researchers call "experience inflation" [3]. In the software and IT sectors, more than 60% of entry-level job listings require three or more years of experience [3].

The sectors that disproportionately employ college graduates—information services, finance, and professional and business services—have reduced hiring more sharply than other industries over the past several years [4]. Meanwhile, the fields showing growth—healthcare, government, and leisure/hospitality, which accounted for nearly 75% of all jobs added in late 2024 and 2025—are not the industries that recent graduates trained for or expected to enter [3].

In the UK, tech graduate roles fell 46% in 2024, with projections of a further 53% decline by 2026 [3]. In India, basic data entry roles that formed the bedrock of the BPO sector are being replaced by multimodal AI at a fraction of the cost [9]. The impact is global and accelerating.

What Comes Next

The gap between the alarm and the policy response remains wide. The Warner-Hawley legislation focuses primarily on data collection and reporting—necessary first steps, but not the kind of structural intervention that a 30%+ youth unemployment rate would demand. The Brookings Institution has cautioned that retraining programs "often struggle to reach those who need them most—individuals with limited education, access to resources, or geographic mobility" [13]. The rapid pace of AI development means that skills taught in retraining programs today may be obsolete before graduates complete them.

ServiceNow's own chief transformation officer has proposed one potential model: hiring early-career "digital natives" into a rotating "problem-solving pool" that circulates through companies until workers find roles matching their skills [2]. It is an acknowledgment that the traditional path from campus to cubicle to career is breaking down—and that even companies driving the disruption recognize the need for new structures.

The question now is whether McDermott's mid-30s prediction proves prescient or merely provocative. The current trajectory—rising graduate unemployment, collapsing entry-level hiring, accelerating AI adoption, and legislative inertia—suggests the answer may depend less on the technology itself than on the speed at which institutions adapt to its consequences. As Senator Warner put it, "If technology eliminates jobs at scale, companies bear responsibility for helping displaced workers transition" [12].

For the class of 2026 and beyond, that transition has already begun—whether they are ready or not.