Airbnb CEO warns: Don't replace entry-level talent with AI
Airbnb CEO Brian Chesky is urging leaders to resist a blunt trade-off: cutting entry-level roles because AI can do the work. His point is simple-if interns don't get reps today, you'll have a leadership vacuum tomorrow.
Yes, automation can take on simpler, lower-level tasks. But organizations still run on human judgment, relationships, and the kind of context you only gain by doing real work over time. As Chesky put it, AI is a tool, not magic-and people will still want relationships and leadership.
Why this matters for executives
- Leadership pipeline risk: Remove entry-level roles and you thin out your future managers. In 3-7 years, you'll be promoting people who never learned the basics.
- Organizational memory: Apprenticeship builds shared standards and culture. Pure automation doesn't pass that down.
- Execution quality: AI handles repeatable tasks. It struggles with ambiguity, cross-team alignment, and trade-offs-the core of leadership.
- Talent market signal: If internships and junior roles dry up, top graduates opt out of your industry-or your company.
Use AI aggressively-still hire juniors
- Ring-fence early-career hiring: Set a floor (for example, 10-15% of hires) that stays intact regardless of automation gains.
- Redesign roles, don't remove them: Let AI handle documentation, first drafts, and routine analyses while juniors own problem scoping, stakeholder comms, and decisions with oversight.
- Apprenticeship model: Pair every manager with 1-2 early-career reports. Make mentoring a scored KPI.
- Structured rotations: 12-18 month rotations across ops, product, and customer-facing teams to build judgment faster.
- Pipeline metrics: Track intern-to-hire rate, time-to-first ownership, manager span of control, and bench strength by function.
- Guardrails for AI: Define what AI may draft vs. what humans must decide. Keep customer conversations, escalations, and strategic planning human-led.
- Tie AI ROI to talent outcomes: Automation is a win only if productivity rises and the bench gets stronger.
What AI should do vs. what humans must do
- AI does: Summaries, first-draft docs, QA checks, data clean-up, meeting notes, basic customer responses, and repetitive ops.
- Humans do: Ambiguous problem solving, negotiation, priority setting, cross-functional alignment, risk calls, and anything that depends on trust.
Signals you're cutting too deep
- Managers spend more time "fixing" AI output than developing people.
- Fewer stretch projects for juniors; more one-off tasks with no learning value.
- Decline in internal mobility within 24 months.
- Customer or partner feedback referencing lack of judgment or empathy.
Quarterly action plan
- Q1: Define AI-use policies by task. Re-open or expand internships. Convert 30-50% of interns to full-time offers where performance merits.
- Q2: Launch a "copilot-first" workflow for junior roles. Set explicit learning goals per role (decision rights, stakeholder mapping, metrics fluency).
- Q3: Implement rotations and a mentorship scoreboard. Publish internal dashboards on pipeline health.
- Q4: Audit promotions for bench depth and diversity. Reinvest AI savings into training and entry-level headcount.
Chesky's warning is pragmatic: even if AI can do the interns' work, future executives still need the mileage that only entry-level roles provide. Strip that away and you'll pay for it when the stakes are higher and the decisions are harder.
If you're formalizing AI skills across roles, consider building a clear upskilling path for juniors and managers. A practical starting point is curating role-based programs that blend AI tools with core business skills: Role-based AI courses.
For broader context on workforce shifts and automation, the World Economic Forum's Future of Jobs research is useful reading: Future of Jobs Report.
Your membership also unlocks:
