The New Data Has Arrived

Since our last coverage on AI's effect on entry-level jobs, two major research drops have shifted the conversation — and neither Wall Street firm is telling the same story as Silicon Valley's executive class.

Goldman Sachs Research, published August 13, 2025, now estimates that AI could displace 6-7% of the U.S. workforce if adoption goes wide. Their team notes the impact is expected to be "transitory." Joseph Briggs, who co-leads Goldman's Global Economics team, argues displaced workers will eventually find new positions as AI creates new roles. Goldman's specific number: unemployment rises by half a percentage point during the transition.

Morgan Stanley Research published its own analysis on April 14, 2026, and it echoes Goldman's cautious optimism — with one very important exception.

Younger Workers Are Taking the Hit First

Morgan Stanley economist Diego Anzoategui points directly at workers aged 22 to 27. Unemployment in that cohort has risen the most since 2023, concentrated in occupations flagged as highly AI-exposed: analysts, accountants, and judicial clerks. These aren't warehouse jobs or retail cashiers. These are college-educated, white-collar entry roles — the exact positions young professionals have historically used to build careers.

Anzoategui notes: "The same technology that automates tasks can also augment workers, increase productivity and boost demand in AI-exposed sectors." Morgan Stanley's own methodology notes that evidence of AI disruption among young workers "becomes weaker" depending on which automation exposure measure is used. The researchers acknowledge the data remains difficult to interpret.

Companies Are Using AI as a Layoff Excuse — HBR Says So Directly

Harvard Business Review, in a piece by Thomas H. Davenport and Laks Srinivasan published January 29, 2026, reports that companies are laying off workers "because of AI's potential — not its performance."

Executives at Ford, Amazon, Salesforce, and JP Morgan Chase have publicly stated white-collar jobs will "soon disappear" due to AI. But HBR's reporting shows the actual productivity data doesn't yet justify the scale of cuts being announced. The layoffs are happening on the expectation of AI capability — not proven results.

When a CEO says "AI made us more efficient," that's a narrative. When researchers say "we can't find broad-based displacement in the macro data yet," those are two different claims.

Box's 13 New Job Types: Real or PR?

The New York Times highlighted Box — a Silicon Valley software company — as a counterexample, noting the firm expects to grow headcount as it creates 13 new AI-specific roles: AI architects, AI solutions managers, and similar positions.

One mid-sized software company creating 13 job categories is not a labor market trend. Box hiring AI architects doesn't rehire the 24-year-old accountant whose entry-level auditing work just got automated. Different people, different skills, different outcomes.

Coverage Gaps

Left-leaning outlets like the NYT have led with the "AI creates jobs" angle — Box's story, the legal sector not collapsing as feared. Business press broadly is taking CEO statements about AI-driven efficiency at face value without demanding supporting data. HBR is one of the few outlets explicitly examining this gap. Goldman and Morgan Stanley's actual research — dense, qualified, honest about uncertainty — is often summarized into soundbites that make things sound either more catastrophic or more benign than the numbers support.

What This Means for Regular People

If you're under 30 and working in finance, law, tech, or any field heavy on computer-based analytical tasks, the data shows you are the most exposed group in the current cycle. Morgan Stanley's April 2026 analysis identifies measurable displacement in this cohort.

The Goldman projection of 6-7% potential workforce displacement is a research institution's base estimate under wide adoption.

If your employer announces layoffs citing AI efficiency gains, the HBR reporting suggests asking for productivity numbers. There's a real possibility the AI isn't performing at the level being claimed.

The transition period Goldman calls "temporary" could last years. Half a percentage point of unemployment sounds small. It isn't small if you're the one unemployed.