AI Automation Risk Table by Karpathy

Andrej Karpathy, the former OpenAI researcher and Tesla AI lead, created a public GitHub repository containing an "AI Automation Risk Table" that systematically ranks professions by their exposure to automation from AI tools. The table assigns risk scores to various jobs based on factors like task complexity, creativity requirements, and physical dexterity. Karpathy took the repository down shortly after publishing it, but developer Josh Kale preserved and documented the analysis, confirming its authenticity through commit history.

The table serves as a provocative, data-driven framework for professionals to assess career vulnerability. It moves beyond speculative discussion by applying a structured scoring methodology to job categories, highlighting which roles face immediate versus long-term disruption from advancing AI capabilities like agents and multimodal models. The rapid deletion sparked discussion about the sensitivity of such analyses and their potential impact on workforce planning and education.

This incident underscores the growing need for concrete, accessible tools to understand AI's labor market impact. While similar analyses exist from research institutions, Karpathy's insider perspective and technical credibility give this particular framework significant weight. The table provides professionals with a starting point for skills assessment and adaptation strategies as automation accelerates across knowledge work and service industries.