Careers in the AI Era: Not that job titles disappear, but that the content of work is restructured
When AI is discussed, you might worry that your job will disappear. But what is actually happening is not that entire job titles vanish, but that tasks (work units) shift to AI, and the center of the remaining work changes.
For example, in the work of writing, AI becomes proficient at starting from zero, and people move toward goal setting, material gathering, editing, verification, and decision-making. In other words, the key for a career strategy is not clinging to your current job title, but redesigning your work with AI as a given.
By Occupation: What changes to AI and what remains as value?
Here we examine representative occupations, split into points of change and growth potential. At least apply to areas close to your own work.
1) Engineer: From a person who writes code to the one who makes specifications runnable
- Changed: Standardized implementations, CRUD, test templates, refactoring proposals, and draft documentation. Acceleration via coding assistance like GitHub Copilot and ChatGPT-based tools.
- Remaining/Growing: Requirements definition, architectural design, security, performance, incident response, data design, reviews. The responsibility to ensure AI outputs operate safely remains with humans.
- Strategy: Not only have it write with prompts, but be able to evaluate and fix. Those who can use unit tests, static analysis, vulnerability assessments (SAST/DAST) in combination are strong.
2) Product Manager (PM): From the person who writes PRDs to the one who raises the quality of decision-making
- Changed: Drafts of PRDs, summaries of user interviews, primary information organization for competitive analysis, release notes creation.
- Remaining/Growing: Which issues to prioritize, trade-offs among business/UX/technical constraints, KPI design, stakeholder coordination.
- Strategy: Use AI to speed up information processing, and devote the saved time to creating the premises for decision-making (hypotheses, validation design, measurement). AI is not a tool to produce conclusions, but a tool to organize decision materials.
3) Marketer: From mass production to people who run verification and learning
- Changed: Drafts for ad copy and landing page copy, outlines for SEO articles, generation of variations of creatives, drafts for social media posts.
- Remaining/Growing: Brand consistency, customer understanding (insights), channel design, A/B test design, ways to improve LTV.
- Strategy: Move from making to testing/validating. As production costs drop with generative AI, the difference shows in the speed of the hypothesis → experimentation → learning cycle, and in primary information (customer voices).
