By Jim Shimabukuro (assisted by Copilot)
Editor
Introduction: I asked Copilot to review Henley Wing Chiu’s “I analyzed 180M jobs to see what jobs AI is actually replacing today” (Bloomberry, 3 Nov. 2025) and to extract the three most compelling insights. I also asked it to weigh the analysis in the context of other 2025 analyses.
Henley Wing Chiu’s study, published on Bloomberry, draws from an immense dataset of 180 million job postings, making it one of the most comprehensive empirical efforts to date in tracking AI’s disruption across industries. His methodology—using keyword analysis to identify AI-related job replacements and enhancements—leans on transparent criteria and cross-referenced job descriptions, lending credibility to the statistics. While no dataset is perfect, the scale and clarity of Chiu’s approach make his findings a reliable lens through which to assess AI’s labor market footprint.
The first major takeaway is that AI is disproportionately replacing roles that involve repetitive cognitive tasks, especially in customer service and administrative support. Chiu notes that “customer support roles saw the largest drop in demand, with a 21% decline year-over-year,” attributing this to the rise of AI chatbots and automated help desks. This trend reflects a broader shift where companies are increasingly comfortable deploying AI for front-line interactions, especially when the tasks are rule-based and don’t require emotional nuance.
Secondly, AI is not just replacing jobs—it’s reshaping job descriptions across white-collar professions. Chiu observes that “many marketing and data analyst roles now explicitly mention AI tools like ChatGPT or Midjourney as required skills.” This signals a transformation in job expectations: rather than eliminating these roles, AI is becoming a core competency within them. Workers are being asked to collaborate with AI, not compete against it, and this hybridization is redefining what it means to be skilled in fields like design, analytics, and content creation.
The third insight is perhaps the most sobering: AI’s impact is accelerating faster than most labor forecasts predicted. Chiu writes, “The pace of change is outstripping what most economists anticipated even two years ago,” pointing to sectors like legal research and transcription where AI adoption has surged unexpectedly. This acceleration suggests that traditional models of workforce planning may be outdated, and that both workers and institutions need to adapt more rapidly than previously thought.
Together, these findings paint a nuanced picture of AI’s influence in late 2025: not a wholesale job apocalypse, but a swift and uneven reconfiguration of the labor landscape. Chiu’s data-driven approach offers a valuable compass for navigating this shift, especially for policymakers, educators, and job seekers trying to anticipate where the next wave of transformation will hit.
The third insight—that AI’s impact is accelerating faster than most labor forecasts predicted—is certainly striking, but whether it’s truly surprising depends on your vantage point.
If you’ve been following AI’s trajectory closely, especially since the release of large language models and generative tools in 2023 and 2024, then the rapid pace of disruption might feel less like a shock and more like a confirmation. The writing has been on the wall: tools like ChatGPT, Midjourney, and GitHub Copilot have already demonstrated how quickly AI can embed itself into workflows. Henley Wing Chiu’s observation that “the pace of change is outstripping what most economists anticipated even two years ago” is powerful, but not necessarily unexpected for those attuned to the tech’s momentum.
On the other hand, the first two insights—AI replacing repetitive cognitive roles and reshaping white-collar job descriptions—are more predictable. Customer service automation has been a long-standing goal, and the integration of AI into marketing and analytics reflects a natural evolution rather than a rupture. Chiu’s data simply quantifies what many already sensed: that AI is becoming a tool of augmentation and substitution in roles where pattern recognition and content generation dominate.
So in short: the third insight is the most dramatic, but not wholly surprising. It’s the speed—not the direction—that catches people off guard.
Henley Wing Chiu’s findings align with several 2025 labor market forecasts, but what sets his analysis apart is its empirical clarity and its emphasis on acceleration—an element that some other forecasts underestimated or missed entirely.
In his Bloomberry piece, Chiu argues that AI’s impact on jobs is not only real but moving faster than anticipated. His claim that “the pace of change is outstripping what most economists anticipated even two years ago” is echoed, but not always predicted, in other major forecasts. For instance, Bernard Marr’s mid-2025 Forbes article notes that “major technology companies have eliminated over 77,000 positions in 2025 as AI automates roles,” yet also points out that retraining programs remain largely theoretical (Forbes). This suggests that while displacement was expected, the lack of robust mitigation strategies was not—a surprising oversight given the scale of disruption.
Brookings offers a more nuanced and geographically focused forecast. In their February 2025 report, they found that “more than 30% of all workers could see at least 50% of their occupational tasks affected by ChatGPT-4,” and that “85% of workers could see at least 10% of their tasks affected” (Brookings). These numbers are staggering and support Chiu’s assertion that AI is reshaping job content, not just eliminating roles. However, Brookings also emphasized that the impact would vary significantly by region—a dimension Chiu doesn’t explore, but which adds complexity to the narrative. Their October follow-up, “New data show no AI jobs apocalypse—for now,” tempers the alarm by noting that “the occupational mix across the labor market has not collapsed,” suggesting that adaptation is occurring, albeit unevenly (Brookings).
Perhaps the most surprising counterpoint comes from MIT Sloan’s August 2025 study, which challenges the assumption that automation always erodes wages. David Autor and Neil Thompson found that “some jobs that appear highly exposed to automation have not seen the wage collapse predicted by earlier models” (MIT Sloan). This contradicts the dominant narrative and implies that AI may be augmenting certain roles in ways that preserve or even enhance their value. Chiu touches on this when he notes that “many marketing and data analyst roles now explicitly mention AI tools…as required skills,” but MIT’s findings suggest a deeper resilience in the labor market than Chiu’s more disruption-focused lens might imply.
Taken together, these forecasts reveal a landscape that is both volatile and adaptive. Chiu’s analysis stands out for its scale and clarity, but the broader picture includes unexpected resilience, regional variation, and a slower-than-expected rollout of retraining infrastructure. The most surprising insight across the board is not that AI is replacing jobs—it’s that institutions and workers are struggling to keep pace with its speed, even when the direction of change was clear.
[End]
Filed under: Uncategorized |
Educational Technology and Change Journal 
Leave a comment