Which Jobs Will AI Replace by 2028? The Projections, the Charts, and the Honest Trajectory

Projecting AI job displacement forward to 2028 requires distinguishing what’s already happened (covered in the 2026 post) from what credible research suggests will happen next. By 2028, frontier models will have crossed additional capability thresholds, agent technology will have matured substantially, and the workforce restructuring that started visibly in 2023-2024 will have compounded. This post walks through the 2028 projections from credible sources, the categories at greatest additional risk, and the honest uncertainty around the numbers.

The 2026 → 2028 delta#

Between mid-2026 and the end of 2028, three structural changes will compound:

1. Agent capability maturation. What requires substantial human oversight today (multi-step autonomous task execution) will be more reliable by 2028, expanding the scope of work that can be AI-delegated.
2. Continued model capability improvement. GPT-6, Claude Opus 5, Gemini 3, plus successors will continue to push the frontier on reasoning, tool use, and reliability.
3. Enterprise adoption deepening. The 2025-2026 enterprise AI deployment has been broad but shallow at many companies. Through 2028 the depth will increase as use cases mature.

The cumulative effect: additional job categories will see substantial displacement, and the already-affected categories will see further reduction.

The categories at greatest 2026-2028 additional risk#

Building on the 2026 baseline, six categories face substantially elevated risk through 2028:

1. Software developers (junior and increasingly mid-level). The 2026 effect has been on entry-level hiring rates. Through 2028, agent-mode coding tools will likely affect mid-level engineering work meaningfully — not by replacing engineers but by making each senior engineer more productive, which compresses the team-size growth that previously required additional hiring. WEF projects software development as a fast-growing role overall but with substantial restructuring of how the growth distributes across seniority.

2. Sales development representatives (SDRs). Outbound sales prospecting is being substantially automated through 2024-2026. By 2028 the SDR role as historically defined — cold email outreach, lead qualification, meeting scheduling — will be largely automated at most B2B companies. Account executives (the seller doing the actual closing) remain.

3. Marketing analysts and digital marketing operators. AI-augmented marketing tooling will compress the workforce required for routine campaign analysis, A/B testing, ad ops, plus content production. Strategic marketing leaders, brand managers, and senior strategists remain.

4. Financial analysts (junior). Routine financial analysis — modeling, valuation comparisons, research synthesis — is increasingly AI-augmented. Senior analysts and investment professionals remain; junior tier hiring is being compressed.

5. HR and recruiting operations. AI-augmented sourcing, screening, and routine HR operations will affect substantial portion of the HR operations workforce. Senior HR business partners, employment lawyers, and HR strategy remain.

6. Radiology and pathology technicians (selective). Increasingly, AI-augmented imaging workflows produce productivity gains that affect technician staffing levels. Radiologists themselves remain in high demand; the technician tier sees compression.

What grows alongside#

The same projections that show contraction in specific categories show growth in others. WEF’s net-positive forecast (170M new jobs vs 92M displaced by 2030) breaks down into specific fast-growing categories:

Fastest-growing roles by 2028	Primary growth driver
AI and machine learning specialists	Direct AI workforce
Data analysts and scientists	AI adoption
Software developers (senior)	Substantial despite junior compression
Cybersecurity analysts	Threat landscape
Renewable energy engineers	Energy transition
Nurses and personal care	Aging populations
Teachers (secondary and specialized)	Demographics
Sales / business development professionals	Volume expansion
Technicians (electrical, robotics, mechanical)	Industrial digitization
Sustainability specialists	ESG and regulation

The pattern is consistent: technical specialists, care professionals, trades, and judgment-based work.

The geographic shifts#

Between 2026 and 2028, geographic shifts will compound:

• India services economy — substantial restructuring as routine BPO compresses but higher-value AI-augmented services grow. The India IT services post covers the dynamics.
• Philippines BPO — particularly exposed; transition to higher-value services is the policy priority.
• Eastern Europe — similar dynamics, transition to specialized services.
• Sub-Saharan Africa — less exposed but also less benefiting; specific programs needed.
• China — restructuring masked by larger demographic and economic dynamics.
• Advanced economies — continued strong demand for AI-fluent workers; transition support for displaced.

The honest uncertainty#

The projections through 2028 carry substantial uncertainty:

Capability uncertainty — AI capability trajectory is not perfectly predictable. The capability that materializes by 2028 will determine the actual displacement.

Adoption pace uncertainty — enterprise adoption depth varies substantially. Organizational and cultural factors slow displacement even when technical capability exists.

Policy response uncertainty — labor regulations, training programs, and political dynamics affect outcomes substantially.

Macroeconomic uncertainty — broader economic conditions affect hiring patterns across all categories.

Productivity-vs-displacement balance — the same AI capability can produce displacement (replacing workers) or augmentation (making workers more productive without replacing). The split depends on managerial choices, organizational culture, and economic conditions.

The honest projection is therefore: the displacement will be real, will be larger than 2026 levels, but the magnitude has substantial uncertainty around it.

What this means for the next two years#

For workers, the implications are concrete:

• Skill development matters more than ever. AI-fluent workers earn premiums. AI-displaced workers face longer transitions.
• Specific categories are at risk; specific categories are growing. Career planning should account for both.
• Reskilling programs are increasingly available — corporate, government, online platforms.
• The transition is real but not apocalyptic — net employment is projected to grow.

For employers, the implications are:

• Productivity gains are substantial but require deliberate adoption.
• Workforce planning needs to account for AI capability.
• Training and reskilling programs are competitive necessities.
• The hiring patterns of 2018-2022 are not the patterns of 2026-2028.

For policymakers, the implications are:

• Transition support for affected workers.
• Skills development at scale.
• Sectoral attention to most-affected categories.
• Economic policy that captures productivity gains broadly.

The 2030 horizon#

For projections beyond 2028, see the 2030 future post.

Where pdpspectra fits#

We help enterprises navigate AI deployment with attention to workforce, productivity, and operational reality. The work spans technical implementation and the organizational change discipline that makes deployment sustainable.

Related reading: the 2026 stats post, the AI agent orchestration post, and the developer productivity metrics post.

---

The 2028 trajectory is bounded but real. Talk to our team about workforce-aware AI deployment.