The Stanford AI Index 2026 is the first report to document AI-driven workforce displacement rather than merely forecast it. We crossed those findings with risk scores for 24 professions to show where the pressure is real, where it is overstated, and what "the jagged frontier" means for your career.
Source: Stanford HAI AI Index 2026
The Stanford AI Index 2026 uses the term "jagged frontier" to describe how AI models can solve extraordinarily complex problems while failing at surprisingly simple ones. Gemini Deep Think earned a gold medal at the International Mathematical Olympiad — yet the top AI model reads analog clocks correctly only 50.1% of the time.
For careers, this means risk is not simply "high-skill = safe, low-skill = at risk." A software developer (rated Low Risk at 38%) saw the sharpest entry-level employment decline of any profession measured. A medical coder (High Risk at 72%) is in the same sector as physicians, who are experiencing dramatic productivity gains from AI — but the type of work differs enormously. You must analyze the task, not just the job title.
Sorted by displacement risk (highest first). Risk scores are composite indices from task routinization, AI tool penetration, and human judgment dependency. Stanford context reflects the 2026 AI Index report and related primary sources.
| Career | Risk Score (%) | Risk Level | Stanford AI Index 2026 Context |
|---|---|---|---|
| Data Entry Clerk | 95 | Critical | Highest automation exposure. WEF projects 85M jobs displaced by 2025; data entry is consistently in the top category. AI achieves 99%+ accuracy on structured data extraction. |
| Bookkeeper | 90 | Critical | QuickBooks AI and similar tools automate up to 80% of reconciliation and categorization tasks. The remaining value is judgment around exceptions and client advisory. |
| Tax Preparer | 88 | Critical | AI now handles form population, deduction identification, and compliance checks for standard returns. Complexity and judgment-heavy filings remain human work. |
| Customer Service Representative | 82 | Critical | Stanford AI Index 2026 documents entry-level employment decline mirroring software development. AI tools deliver 14–15% productivity gains; routine ticket resolution is heavily automated. |
| Administrative Assistant | 82 | High Risk | See full career analysis for details. |
| Accountant | 80 | Critical | Automation covers auditing, reconciliation, and reporting. Stanford AI Index 2026 notes AI penetration in finance operations is among the highest across sectors. |
| Paralegal | 78 | High Risk | LawGeex and contract-review AI have automated document review, due diligence, and research tasks. Stanford AI Index 2026 data shows legal services in top-five AI-exposed categories. |
| Content Writer | 76 | High Risk | See full career analysis for details. |
| Insurance Underwriter | 74 | High Risk | Algorithmic underwriting now handles standard risk assessment with higher accuracy than manual review. Human underwriters increasingly focus on complex commercial lines. |
| Medical Coder | 72 | High Risk | AI diagnostic coding tools process clinical notes with 90%+ accuracy. Physicians using AI note-generation tools report 83% less time on documentation (Stanford AI Index 2026). |
| Graphic Designer | 72 | High Risk | See full career analysis for details. |
| Loan Officer | 65 | Moderate | Algorithmic credit decisioning has largely automated consumer loan origination. Commercial lending and relationship banking remain resistant to full automation. |
| IT Support Specialist | 65 | Moderate | AI helpdesk systems resolve Tier-1 and many Tier-2 tickets without human intervention. However, complex infrastructure and security incidents still require human expertise. |
| Real Estate Agent | 65 | High Risk | See full career analysis for details. |
| Sales Representative | 58 | Moderate | CRM AI automates lead scoring, outreach sequencing, and pipeline forecasting. Top performers differentiate through relationship depth and complex deal navigation. |
| HR Manager | 58 | Moderate | See full career analysis for details. |
| Financial Analyst | 55 | Moderate | Bloomberg AI and similar tools automate data gathering, model construction, and report generation. Interpretation, client communication, and novel situations remain human work. |
| Data Analyst | 45 | Moderate | AI co-pilots now write SQL, generate visualizations, and surface anomalies automatically. Stanford AI Index 2026 notes AI skills demand in the information sector grew from 7.8% to 13.2% in 2025. |
| Investment Banker | 42 | Moderate | AI handles M&A due diligence, comparable analysis, and pitch deck drafting. Senior bankers find their value in relationships and deal structuring judgment, which are highly resistant to automation. |
| Software Developer | 38 | Low Risk | Entry-level developer employment (ages 22–25) fell ~20% since 2024 per Stanford AI Index 2026, even as senior headcount grew. AI delivers +26% productivity in dev tasks. |
| Teacher | 35 | Low Risk | See full career analysis for details. |
| Product Manager | 32 | Low Risk | AI accelerates user research synthesis, roadmap generation, and A/B analysis. Core PM value — stakeholder alignment, strategic judgment, and cross-functional leadership — remains human-dependent. |
| Cybersecurity Analyst | 28 | Low Risk | A dual-edged data point: AI security agents achieved 93% problem-solving success vs. 15% in 2024 (Stanford AI Index 2026). AI augments analysts rather than replacing them, as threat complexity grows in parallel. |
| Registered Nurse | 22 | Low Risk | See full career analysis for details. |
Highest automation exposure. WEF projects 85M jobs displaced by 2025; data entry is consistently in the top category. AI achieves 99%+ accuracy on structured data extraction.
QuickBooks AI and similar tools automate up to 80% of reconciliation and categorization tasks. The remaining value is judgment around exceptions and client advisory.
AI now handles form population, deduction identification, and compliance checks for standard returns. Complexity and judgment-heavy filings remain human work.
Stanford AI Index 2026 documents entry-level employment decline mirroring software development. AI tools deliver 14–15% productivity gains; routine ticket resolution is heavily automated.
Automation covers auditing, reconciliation, and reporting. Stanford AI Index 2026 notes AI penetration in finance operations is among the highest across sectors.
LawGeex and contract-review AI have automated document review, due diligence, and research tasks. Stanford AI Index 2026 data shows legal services in top-five AI-exposed categories.
Algorithmic underwriting now handles standard risk assessment with higher accuracy than manual review. Human underwriters increasingly focus on complex commercial lines.
AI diagnostic coding tools process clinical notes with 90%+ accuracy. Physicians using AI note-generation tools report 83% less time on documentation (Stanford AI Index 2026).
Algorithmic credit decisioning has largely automated consumer loan origination. Commercial lending and relationship banking remain resistant to full automation.
AI helpdesk systems resolve Tier-1 and many Tier-2 tickets without human intervention. However, complex infrastructure and security incidents still require human expertise.
CRM AI automates lead scoring, outreach sequencing, and pipeline forecasting. Top performers differentiate through relationship depth and complex deal navigation.
Bloomberg AI and similar tools automate data gathering, model construction, and report generation. Interpretation, client communication, and novel situations remain human work.
AI co-pilots now write SQL, generate visualizations, and surface anomalies automatically. Stanford AI Index 2026 notes AI skills demand in the information sector grew from 7.8% to 13.2% in 2025.
AI handles M&A due diligence, comparable analysis, and pitch deck drafting. Senior bankers find their value in relationships and deal structuring judgment, which are highly resistant to automation.
Entry-level developer employment (ages 22–25) fell ~20% since 2024 per Stanford AI Index 2026, even as senior headcount grew. AI delivers +26% productivity in dev tasks.
AI accelerates user research synthesis, roadmap generation, and A/B analysis. Core PM value — stakeholder alignment, strategic judgment, and cross-functional leadership — remains human-dependent.
A dual-edged data point: AI security agents achieved 93% problem-solving success vs. 15% in 2024 (Stanford AI Index 2026). AI augments analysts rather than replacing them, as threat complexity grows in parallel.
Methodology: scores are composite of task routinization (40%), AI tool penetration (35%), and human judgment dependency (25%). Based on Oxford/Frey & Osborne, McKinsey Global Institute, WEF Future of Jobs 2025, and Stanford HAI AI Index 2026. Scores represent baseline risk without AI augmentation skills. Full methodology →
The 2026 Stanford AI Index is the first edition to document concrete AI-driven workforce displacement rather than merely forecast it. Employment among US software developers aged 22–25 fell nearly 20% since 2024, even as senior developer headcount grew. One-third of organizations expect AI to reduce their workforce in the coming year, with highest anticipated cuts in service operations, supply chain, and software engineering.
The "jagged frontier" describes how AI models can excel at extraordinarily complex tasks while failing at surprisingly simple ones. For example, Gemini Deep Think earned a gold medal at the International Mathematical Olympiad, yet top AI models read analog clocks correctly only 50.1% of the time. This means the risk profile of AI for any given job is uneven — some tasks are fully automated while others remain human-dependent in non-obvious ways.
Based on our composite risk model (task routinization 40%, AI tool penetration 35%, human judgment dependency 25%), the five highest-risk professions are: Data Entry Clerk (95%), Bookkeeper (90%), Tax Preparer (88%), Customer Service Representative (82%), and Accountant (80%). These align with Stanford AI Index 2026 data showing the largest productivity gains and employment shifts in exactly these high-task-routinization roles.
According to the Stanford AI Index 2026, productivity gains from AI vary significantly by field: customer support sees 14–15% gains, software development sees 26% gains, and marketing output improves up to 50% on targeted tasks. However, for tasks requiring judgment, reasoning, or novel problem-solving, the gains are smaller or inconsistent. Physicians using AI note-generation tools report up to 83% less time writing clinical notes.
Mid-career and senior software developers remain relatively safe (our model scores the role at 38% risk), but the Stanford AI Index 2026 reveals a critical caveat: entry-level developer employment for workers aged 22–25 has fallen nearly 20% since 2024. AI delivers 26% productivity gains in software development, meaning organizations can accomplish more with fewer junior developers. The risk is concentrated at career entry, not at the senior level.
Cybersecurity is a dual-edged case. Stanford AI Index 2026 reports that AI security agents achieved 93% problem-solving success on cybersecurity benchmarks, up from 15% in 2024. This dramatically increases analyst productivity but does not eliminate the role — as AI capabilities grow, so does the sophistication of AI-powered attacks, requiring human oversight. Our risk model scores cybersecurity analysts at 28% (Low Risk).
Each score is a composite of three dimensions: Task Routinization (40% weight) — how predictable and rule-based the daily tasks are; AI Tool Penetration (35%) — whether enterprise AI tools already exist and are being adopted for the role; and Human Judgment Dependency (25%) — how much the role relies on contextual reasoning, emotional intelligence, or novel problem-solving. Sources include Oxford/Frey & Osborne (2013), McKinsey Global Institute (2024), WEF Future of Jobs Report (2025), and Stanford HAI AI Index (2026).
The table above shows baseline risk. Your personal risk depends on which tasks within your role you focus on and which AI-adjacent skills you've built. Use the calculator to get your adjusted score.
Calculate My Risk Score →