With more realistic data on the impact of AI on the job market economy, the situation is far from clear.
On the one hand, you have companies like Duolingo announce bold plans to become “AI-First”, and CEO Luis von Ahn announces that language learning apps will “gradually stop using contractors to do what AI can handle.”
this Originated in 2024When the contractors went to Reddit to protest, they received emails about being eliminated.
A commentator at the time shared an email from Duolingo, noting: “This is the last email I received two weeks ago. Just in case you want to see it. I’ve been there for five years. Our team has four core members and we both got the boots. These two will just wait to see what the AI is about to make sure it’s acceptable.”
As Duolingo worked harder on “AI-Fir,” in a note shared with LinkedIn, Von Ahn labeled it as a plan to remove bottlenecks and freelance employees to focus on creative work rather than to directly replace humans.
This is the stance of other tech leaders, such as Shopify CEO Tobi Lütke, who recently told Teams that they need to justify new employees by showing “why can’t use AI to get what they want to do”.
This means that if a task can be automated, it should be – human characters turn to advanced work that machines cannot easily replicate.
However, these “high-level” characters are still somewhat vague and underdefined. Whether AI can match the quality of the original work. Von Ahn’s notes even admit that for the sake of speed, the quality is worth the sacrifice.
on the other hand, New paper Some cold water has poured out from economists Anders Humlum and Emilie Vestergaard that AI is already changing the job market.
Analyzing data from 25,000 workers from more than 25,000 “AI exposed” occupations, such as software developers, journalists and accountants, Humlum and Vestergaard, found that adopting AI Chatbots “has no significant impact on the income or time recorded in any occupation.”
Although the technology is absorbed very quickly, most workers now use AI tools regularly in exposed areas.
Then what to give? According to Humlum, although many workers save time in using AI, these benefits have not yet translated into expanded output or higher income.
New tasks related to AI (such as timely engineering or output quality control) seem to have some productivity gains offset.
In other words, at least so far, AI has not necessarily reduced the demand for labor. It just changes the nature of the work. Of course, it’s still early days, and Humlum quickly noticed that his findings are more like the “upper cap” of short-term impact than predictions of things in the long run.
Indeed, even if some companies talk about becoming a big game of “AI-First”, local reality can be more confusing and more progressive.
As far as Duolingo is concerned, Duolingo emphasizes that the change in its attitude is not about replacing employees (apparently many people regard it as an empty company PR master), but rather supports them through “more training, mentoring and AI tools.” While Shopify is putting responsibility on the team to justify employees, it has not announced any significant layoffs related to automation.
So, what about the place for us? One word: Unsure.
Obviously, AI is expected to become a huge destructive force in the workplace. However, the path from potential to large-scale impact is unlikely to be a straight line.
