Some links on this page are affiliate links: if you buy through them we may earn a commission, at no extra cost to you.
A Danish study of workers using AI chatbots found no measurable average change in earnings or recorded working hours during its study period. That is not the same as finding that AI had no effect: adoption spread, workers reported modest time savings, and tasks changed. The clearest takeaway is that workplace transformation showed up before broad changes in pay or hours—and the findings do not prove that no one lost work or that future AI will have the same effects.
Which study is behind the headline?
The research is by Anders Humlum of the University of Chicago Booth School of Business and Emilie Vestergaard of the University of Copenhagen. Its earlier working-paper title was Large Language Models, Small Labor Market Effects. The latest identified version is NBER Working Paper No. 33777, titled Still Waters, Rapid Currents: Early Labor Market Transformation under Generative AI, revised in March 2026.
The study examines AI-chatbot adoption in Denmark, not every form of artificial intelligence. It combines surveys conducted in late 2023 and 2024 with administrative labor-market records, covering about 25,000 workers across 7,000 workplaces in 11 occupations considered exposed to chatbots. Those occupations included accountants, customer-support specialists, financial advisers, human-resources professionals, IT-support specialists, journalists, legal professionals, marketing professionals, office clerks, software developers and teachers.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
The researchers used differences in employer AI policies as quasi-experimental variation in a difference-in-differences analysis. That approach helps estimate the relationship between adoption and outcomes, but it does not turn a short, country-specific study into a definitive forecast for all workers.
#1 Best Overall
What changed—and what did not?
| Evidence of change | No detectable broad change in the study window |
|---|---|
| AI chatbot adoption and employer initiatives spread quickly. | Average earnings did not show a statistically significant change attributable to adoption. |
| Workers reported modest time savings; an earlier paper version estimated about 3% on average. | Recorded working hours did not show a statistically significant change. |
| Work involved new tasks such as generating content, checking and editing outputs, overseeing AI, and integrating it into workflows. | The study did not find broad occupation-level displacement in the outcomes it measured. |
| Some adopters moved toward higher-paying occupations where chatbots were more relevant, but the number was too small to shift average earnings. | The aggregate results do not establish that every worker, employer or occupation was unaffected. |
The latest version reports that its estimates rule out effects larger than approximately 2% on earnings and recorded hours two years after ChatGPT’s launch. An earlier version used a different estimate, describing confidence intervals that ruled out effects larger than 1%. These figures come from different revisions and should not be combined as if they were one unchanged result. Likewise, the roughly 3% time-saving estimate belongs to the earlier paper version; it is not a finding that wages or total firm productivity rose by 3%. See the earlier working-paper version and the latest NBER record.
How can AI save time without raising wages?
Several steps separate completing a task faster from receiving higher pay. Time saved on a draft or summary may be spent on more assignments, not on leisure. A worker or manager may need to check facts, revise style, handle compliance, or fit the output into existing systems. Even when a tool improves a task, the gain may be too small to change a workplace’s overall output or finances.
And a gain in productivity does not automatically become a wage increase. Whether workers receive more depends on factors such as the scale of the gain, employers’ decisions, labor demand and workers’ bargaining position. The study’s results are consistent with modest time savings and task reorganization without measurable wage pass-through in the observed period; they do not prove which explanation accounts for every workplace.
Free tools Windows power users keep installed
One-click scans. No signup required.
This also helps reconcile workplace studies with controlled experiments that find larger gains on specific tasks. A test may give someone a clearly defined assignment and measure its completion. In an actual job, assignments shift, information may be restricted, outputs need review, and coordination and approval take time. Faster performance on one task does not by itself show that a firm can remove a position or raise every employee’s pay.
Does this mean ChatGPT is not replacing jobs?
Only in a limited sense. The study found no broad reduction in recorded hours or earnings attributable to chatbot adoption in the selected Danish occupations during the period examined. That is evidence against rapid, widespread displacement showing up in those measures at that time. It does not show that no individual lost a job, that employers did not slow hiring, or that freelancers and contractors were unaffected.
Jobs can also change without disappearing. AI may take over part of a task while workers take on review, verification, workflow design or other work. A company may use a tool to handle more business without hiring additional staff, or reduce future recruitment instead of dismissing current employees. Average results can conceal gains for some workers and losses for others. Recorded hours and earnings do not capture every change in workload, autonomy, job quality or advancement.
The distinction matters especially for early-career work. The latest paper’s summary says its null results also hold for early-career jobs. That does not settle whether junior writing, research or support tasks will shrink over time, or whether there will be fewer opportunities to learn those tasks on the job. Hiring criteria, training assignments and promotion paths may change before broad employment totals do.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Why Denmark’s results may not transfer directly to other countries
The evidence comes from Denmark, whose labor-market institutions, collective bargaining, employment protections, occupational mix and welfare system differ from those in the United States and elsewhere. The study is informative about early chatbot adoption in its setting; it cannot establish that American workers experienced the same effects.
Its observation window is also short relative to major technological shifts. The researchers studied 11 selected occupations and chatbot use, not the whole economy, industrial robots or every kind of automation. Average effects may hide concentrated impacts on particular firms, tasks or groups. Administrative earnings and hours are valuable measures, but they cannot by themselves reveal changes in hiring freezes, job quality, career ladders or work intensity.
Rank #4
- ENHANCED CONTEXT WITH MULTIMODAL INPUT: Capture audio, type notes, add images, and press to highlight key moments for richer context. During recording, instantly mark key moments with a single button press. Simultaneously enrich your audio by snapping photos of important documents or typing in ideas
- CHAT WITH YOUR RECORDINGS USING "ASK Plaud": Unlock deeper insights with this interactive AI. Ask questions, extract key points, draft emails, and get next-step suggestions—all grounded in your original audio for reliable, ready-to-use answers
- INTELLIGENT RECORDING WITH AI DIRECTIONAL AUDIO: Enjoy seamless, intelligent recording with Plaud Note Pro. Its AI automatically switches between call and meeting modes while recording, while directional audio and real-time spatial awareness minimize noise to capture voices with crystal clarity
- Everything Included: Includes Plaud Note Pro, magnetic case, magnetic ring, charging cable, and a free Starter Plan with 300 transcription minutes per month. Upgrade anytime in the Plaud app to Pro Plan (1,200 min/mo) or Unlimited Plan(Up to 24 hours of transcription per user per day)
- PREMIUM ULTRA-SLIM DESIGN WITH INSTANTVIEW DISPLAY: Meticulously designed, the AI Note Taker is just 0.12 inches thin and 1.06 oz —about the size of a credit card. Its sleek aluminum body with a textured wave finish features a vivid AMOLED display, letting you check battery and recording status at a glance, while it seamlessly works with Apple Find My to ensure you never misplace it
The paper is a working paper, so its findings should be treated as research evidence rather than a final consensus. Future systems could have different effects, particularly if companies deploy AI agents that take actions across software, use proprietary business information, or complete multi-step workflows. Such changes could prompt firms to redesign departments rather than simply assist individual tasks. That is a possibility, not a prediction established by this study.
What workers and employers should watch next
To judge whether AI is changing work beyond experimentation, look beyond headlines about adoption. Useful signals include entry-level hiring and job postings, earnings by experience level, hours and workload intensity, layoffs and redeployment, internal moves between occupations, and whether job descriptions add AI-related duties. It also matters whether saved time becomes more output, more free time, reduced staffing or some combination.
PC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchFor an individual worker, learning to use a chatbot may be useful, but this study does not show that buying a particular AI subscription protects a job or guarantees a raise. Employers evaluating tools should measure the whole workflow—including review time, error correction, data access and security—not just how quickly a model produces a first draft. The key question is whether the tool improves work after those costs are included.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
