I’ve heard endless predictions about how AI chatbots work, but the data draws a much more subdued picture for at least now.
Despite the enormous and continuous advances in generative AI, the enormous waves that it was supposed to create in the world it created so far look like a ripple.
Researchers Anders Humram (University of Chicago) and Emily Westergaard (University of Copenhagen) did not rely on anecdotes. They dig deeper and linked responses from two major surveys (late 2023 and 2024) with official detailed records of employment and salaries in Denmark.
The pair zoomed in on around 25,000 people working in 7,000 different locations, covering 11 jobs that were thought to be on the path of AI’s chaos.
Everyone uses AI chatbots for their work, but where are the benefits?
What they find checks what many of us see: AI chatbots are currently everywhere in our workplaces in Denmark. Most bosses actually encourage staff to use them. This is a real shift from the early days when companies were naturally nervous about things like data privacy.
Nearly four of the 10 employers have their own internal chatbots, and nearly a third of their employees have been formalized training in these tools.
When the boss nodded, the number of staff using the chatbot effectively doubled, jumping from 47% to 83%. It also helped to level the playing field a little. What is the gap between men and women using chatbots? It has been significantly reduced when companies actively encouraged their use, especially when they received some training.
So, tools are popular, businesses are investing, people are trained…but is there a big economic change? It appears to be missing while working.
Researchers basically had nothing to do with people who used statistical methods to use AI chatbots to work with people who didn’t explode into the scene before and after ChatGpt.
“An exact zero,” researchers call their findings. There were no major conflicts in wages, and changes in recorded working hours across all 11 job types they saw. And they are pretty confident in this – the numbers rule out average effects greater than 1%.
This wasn’t just a blip either. The lack of shock was also true for those who boarded early, those who use chatbots every day, or those who work in places where their bosses are actively pushing their skills.
Looking at the whole workplace, the story didn’t change. Where many chatbot users were present, we saw no different trends in employment, overall wages, or retention compared to lower employment, overall wages, or staff use.
Increased productivity: More gentle nudges than protruding
Why is there a big cut? Why are there all the hype and investments when they don’t appear in your salary or job statistics? The study flags two main perpetrators. Increased productivity is not as great as it is expected in the real world.
Certainly, those who use AI chatbots for their work found them useful. They mentioned better quality of work and more creative feelings. But what is the most advantageous? Save time.
However, when researchers calculated the numbers, the average time saved was only about 2.8% of the total working hours of users. This leaves the huge 15%, 30%, and even 50% productivity jumps seen in controlled lab-style experiments (RCTs) that involve similar work.
Why the difference? It seems a few things are happening. These experiments focus on jobs and specific tasks that chatbots really shine through (such as coding help and basic customer service responses). This study examined a wider range, including jobs such as education, which could lead to lower benefits.
Researchers emphasize the importance of what is called “complementary investments.” Those who encouraged and provided training by businesses to use chatbots have actually reported greater benefits. Save more time, improve quality and feel more creative. This suggests that having a tool isn’t enough. To truly unlock that possibility, you need the right support and the company environment.
And even these modest time savings didn’t pad your wallet. This study believes that only a small portion of the time saved (probably 3% to 7%) actually manifested as higher revenues. It could be standard workplace inertia, and it could be difficult to ask for a salary increase, especially when many people start using them from their own bats, based on the use of tools their bosses aren’t officially blessed.
Doing a new job, not that much work
One fascinating twist is that AI chatbots don’t just perform old work tasks faster. It appears they are also creating new tasks. Approximately 17% of those using them said they had new workloads, most of them had new types of tasks.
This phenomenon occurred more frequently in workplaces that promoted the use of chatbots. It spilled on people who were not using the tool. Approximately 5% of non-users reported that new tasks pop up because they have to adjust their AI, particularly assignments, or find homework they wrote.
What new task? It involves figuring out how to weave AI into your daily workflow, thinking about how to use AI help to draft content, and, importantly, addressing ethical aspects, making sure everything is on the board. It suggests that companies are still spending so much time and effort at the stage of “getting it out” and not enjoying immediate rewards.
What is the verdict regarding the impact of AI chatbots on their work?
Researchers are careful not to completely amortize the generated AI. They see a pathway that becomes more influential over time, especially as companies become better at consolidating it, and perhaps as those “new tasks” evolve.
But for now, their message is clear. The current reality is not consistent with the hype about large, immediate job market overhauls.
“Despite rapid recruitment and substantial investment, our key findings are that AI chatbots have minimal impact on productivity and labor market outcomes so far,” the researchers conclude.
It reminds us that it can be seen everywhere except for the old quotes about the early computer age: productivity statistics. Two years after ChatGpt launched, we have begun adopting the fastest technology we’ve ever seen. The actual marks on jobs and wages seem surprisingly light.
The revolution may still be coming, but it seems to be taking time.
See: Claude Integrations: Humanity adds AI to his favorite work tools
Want to learn more about AI and big data from industry leaders? Check out the AI & Big Data Expo in Amsterdam, California and London. The comprehensive event will be held in collaboration with other major events, including the Intelligent Automation Conference, Blockx, Digital Transformation Week, and Cyber Security & Cloud Expo.
Check out other upcoming Enterprise Technology events and webinars with TechForge here.
