Will AI replace my job?

It depends on your occupation and specific tasks. Computer programmers (75% task coverage), customer service reps (70%), data entry keyers (67%), telemarketers (65%), and proofreaders (63%) face the highest AI displacement risk according to Anthropic's 2026 research. Physical roles like electricians and massage therapists score under 10.

Which jobs are most at risk from AI automation?

Computer Programmers 75%, Customer Service Reps 70%, Data Entry Keyers 67%, Telemarketers 65%, Proofreaders 63%, Tax Preparers 62%, Financial Analysts 60%, Paralegals 58%, Medical Transcriptionists 57%, Bookkeeping Clerks 55%.

Is AI causing unemployment?

As of early 2026, no statistically significant unemployment increase has been found for AI-exposed workers. However, a 14% drop in new hiring for workers aged 22–25 entering high-exposure occupations has been documented post-ChatGPT.

How is the AI job risk score calculated?

The score combines the O*NET task database (800+ US occupations), the Anthropic Economic Index (real Claude professional usage data from Aug/Nov 2025), and Eloundou et al. beta scores. Automated use is weighted 1.0x and augmentative use 0.5x.

ANTHROPIC RESEARCH · MARCH 2026

Will AI Replace Your Job?

Research-backed risk scores for 800+ occupations, grounded in Anthropic's 2026 labor market study.

800+

Occupations Analyzed

75%

Highest Task Coverage

14%

Hiring Drop (Ages 22–25)

47%

Higher Pay in Exposed Jobs

How It Works

🗂

O*NET Task Database

800+ US occupations broken into specific tasks with time-weight estimates from the Department of Labor.

📡

Real LLM Usage Data

Anthropic Economic Index: actual Claude usage patterns from professional conversations, August and November 2025.

Eloundou β Scores

Theoretical LLM feasibility scores (β=1.0: 2× faster alone; β=0.5: needs tools; β=0.0: not feasible).

Key Research Findings

📉

No Unemployment Spike (Yet)

No statistically significant unemployment increase detected for even the most AI-exposed workers as of early 2026.

👶

Entry Level Hit Hardest

Workers aged 22–25 entering high-exposure occupations show a 14% drop in job-finding rate post-ChatGPT.

💰

High Earners at Risk

Exposed workers earn 47% more on average. Graduate degrees are 3.9× more common — AI flips the historical disruption pattern.

BLS Confirms the Signal

Every 10 percentage points of observed AI exposure correlates with 0.6pp lower projected employment growth (2024–2034).

🔬

Observed, Not Theoretical

Unlike prior models, this uses real Claude usage data from professional settings, not just theoretical task feasibility.

🔗

Task Bundling Matters

60% task exposure doesn't automate a job if the remaining 40% requires physical presence, legal accountability, or emotional intelligence.

Find Out Your Risk Score

Get an instant AI exposure analysis for any job title — backed by Anthropic's 2026 labor market research. Free, no signup required.