AI Safety Researcher vs Penetration Tester: Which Pays More?
Side-by-side salary comparison by city, experience level, and career growth outlook. Data reflects current market rates.
AI Safety Researcher
Research methods to make AI systems safer, more aligned, and interpretable.
Penetration Tester
Conduct authorised attacks on systems to find and report vulnerabilities.
Penetration Tester earns more on average — the national median is $47,690/year (59%) higher than a AI Safety Researcher. However, salaries vary significantly by city, employer, and experience level — see the city-by-city breakdown below.
AI Safety Researcher vs Penetration Tester — Salary by City
National median figures in USD across top cities.
| City | AI Safety Researcher | Penetration Tester | Difference |
|---|---|---|---|
| San Francisco, CA | $118,871 | $193,496 | AI $74,625 |
| New York, NY | $79,460 | $140,470 | AI $61,010 |
| Seattle, WA | $113,639 | $188,841 | AI $75,202 |
| Austin, TX | $92,366 | $145,769 | AI $53,403 |
| Chicago, IL | $85,872 | $135,251 | AI $49,379 |
| Boston, MA | $107,617 | $171,501 | AI $63,884 |
| London, UK | £63,159 | £102,480 | AI $39,321 |
| Toronto, Canada | CA$111,112 | CA$166,911 | AI $55,799 |
Frequently Asked Questions
Does a AI Safety Researcher or Penetration Tester earn more?+
A Penetration Tester earns more on average. The national median salary for a AI Safety Researcher is $81,490/year, compared to $129,180/year for a Penetration Tester — a difference of $47,690 (59%).
Which has better career growth — AI Safety Researcher or Penetration Tester?+
AI Safety Researcher roles are growing at 65% YoY while Penetration Tester demand is growing at 30% YoY. AI Safety Researcher has stronger near-term demand growth.
Can you switch from AI Safety Researcher to Penetration Tester?+
Yes. Many professionals transition between these roles, especially since both are in the same category. Shared skills include: Python.
Which is harder to automate — AI Safety Researcher or Penetration Tester?+
AI Safety Researcher has a lower AI automation risk (8% vs 22%). Based on Oxford Martin School and McKinsey 2023 analysis.