Here are four key takeaways from the article:
- AI chatbots saved workers just 3% of their time, with only 3–7% of those gains reflected in pay increases.
- The technology had no significant impact on earnings or recorded hours across any occupation studied.
- Most saved time was spent on other work tasks, including new responsibilities created by AI use.
- Lack of employer support, training, and organizational change limits AI’s workplace impact.
AI chatbots were supposed to transform work as we know it. But two years into the generative AI boom, new research shows the real impact may be far less exciting.
A working paper from the National Bureau of Economic Research (NBER) reveals that AI chatbots—despite being rapidly adopted across white-collar jobs have had a negligible effect on employee pay or working hours. The study tracked over 25,000 workers in Denmark across 7,000 workplaces and found only a 3% time savings on average, with just 3% to 7% of productivity gains making their way into paychecks.
Table of contents
High Expectations, Low Returns
Since the launch of ChatGPT, companies have been racing to integrate AI chatbots in the workplace, hoping to boost workplace productivity and cut costs. But the new data tells a different story. According to economists Anders Humlum and Emilie Vestergaard, most workers saw little benefit in terms of extra income or reduced work hours.
Even jobs believed to be highly susceptible to AI disruption like software development, marketing, and HR showed limited productivity boosts and minimal compensation increases.
Productivity Gains Aren’t Tricking Down
One surprising takeaway is that while AI helped some workers complete tasks faster, this didn’t translate into more free time or higher pay. Instead, over 80% of saved time was spent doing new tasks created by AI tools, like editing AI-generated content or monitoring students to prevent cheating.
So, what’s stopping AI from delivering on its promises? The study suggests a few key reasons:
- Lack of employer buy-in
- Poor time reallocation
- Minimal training and support
- Workers afraid to admit AI is helping them work faster
In other words, while AI tools are available, the organizational structure and culture haven’t caught up.
AI Job Impact May Take Longer to Show
The findings echo growing skepticism in the tech and business world. IBM found that only 25% of AI projects deliver ROI, and companies like Klarna have quietly scaled back bold claims about AI replacing human workers.
Even Nobel laureate Daron Acemoglu estimates that AI will only boost U.S. GDP by 1.1% to 1.6% over the next decade far less than the tech utopia some predicted.
Getting real benefits from AI adoption in companies will require more than plugging in a chatbot. It demands training, trust, and strategic integration into existing workflows. Without that, the AI revolution might stay stuck in neutral.
Final Thoughts: A Long Road Ahead
The AI hype train is still moving fast but this study is a reminder that technology alone isn’t enough. Just like electricity needed decades to power modern assembly lines, AI’s real workplace transformation could still be years away.
Until then, workers may be smarter, but not necessarily richer.
| Latest From Us
- Robotaxis Are Watching You: How Autonomous Cars Are Fueling a New Era of Surveillance
- AI Unmasks JFK Files: Tulsi Gabbard Uses Artificial Intelligence to Classify Top Secrets
- FDA’s Shocking AI Plan to Approve Drugs Faster Sparks Controversy
- AI in Consulting: McKinsey’s Lilli Makes Entry-Level Jobs Obsolete
- AI Job Losses Could Trigger a Global Recession, Klarna CEO Warns
