AI Threatens Jobs that Can Be ‘Unbundled’

Why It Matters

Understanding which roles are unbundlable helps firms prioritize reskilling and informs policymakers about sectors likely to experience the sharpest employment shocks from generative AI.

Key Takeaways

• AI risks jobs with separable tasks
• Strong-bundle roles retain human involvement
• Writers, programmers, designers face highest AI displacement
• Coordination costs shield certain occupations
• Income losses concentrate on analysts and developers

Pulse Analysis

The concept of "unbundling" reframes the AI‑job displacement debate. Rather than treating occupations as monolithic blocks, researchers examine how easily a role’s component tasks can be isolated and handed off to machines. When a task can be performed independently—such as drafting text, writing code, or creating UI mockups—AI tools can replace the human element with minimal friction. Conversely, jobs that require continuous coordination, shared liability, or contextual awareness—think of a project manager overseeing cross‑functional teams—incur high “bundle‑breaking” costs, preserving the human presence even as AI augments performance.

Empirical evidence underscores the theory. Tufts University’s Digital Planet study projects that 57 % of writing and authoring positions, 55 % of computer programmers, and a similar share of web designers could be substantially automated. The same analysis flags software developers, management analysts and market research analysts as the occupations facing the greatest aggregate income erosion, reflecting both wage compression and reduced demand for routine outputs. These figures translate into billions of dollars of potential earnings at risk, reshaping labor market dynamics across the tech sector and creative industries.

For businesses, the takeaway is clear: strategic workforce planning must differentiate between unbundlable and strongly bundled roles. Companies should invest in upskilling employees whose tasks are prone to automation—emphasizing creative problem‑solving, cross‑disciplinary collaboration, and oversight of AI systems. Policymakers, meanwhile, can target safety‑net programs and training grants toward the most exposed occupations. By aligning talent development with the nuanced realities of task unbundling, firms can harness AI’s productivity gains while mitigating disruptive employment impacts.