Last hired, first fired? AI developers may automate themselves out of jobs first

A new perspective on AI automation suggests that novel AI development itself could become the first fully-automated job rather than the last, challenging conventional thinking about which professions will be displaced by artificial intelligence.

The conventional wisdom: Most experts have long assumed AI development would be among the last jobs to be fully automated, since AI systems are needed to automate other professions first.

The contrarian argument: Current data-hungry AI methods may actually make AI development a prime candidate for early automation due to several unique factors.

• AI researchers actively contribute to automating their own field, unlike workers in other industries who may resist or slow down automation efforts.
• The AI development community accepts the inherent risk of their skills becoming automated as part of their professional reality.
• AI developers possess the deepest familiarity with their own field, naturally directing the most concentrated automation efforts toward AI development itself.

Why other jobs might take longer: Traditional industries face significant barriers that could delay their automation compared to AI development.

• Data collection for automating other fields may require lengthy timeframes, even with automated systems gathering information.
• Workers in non-AI fields who fear job displacement may actively stall or resist providing the data and collaboration necessary for their own automation.
• Cross-industry automation requires AI systems to work with potentially uncooperative human experts in each respective field.

The big picture: This analysis suggests that the very people building AI systems may be engineering their own professional obsolescence faster than they’re automating other industries, creating an unexpected sequence in the automation timeline.