Good speculation, but in my opinion, a bit over the top, at least here in Finland.
Yeah, in theory, AI could handle orders in a bookstore, for example, and a robot could even pack them, and the logistics chain could probably be streamlined too.
But what about in practice, when we’re talking about, say, the manufacturing industry? You don’t just suddenly teach AI about assembly—which part of a piece needs to be protected during sandblasting, whether it’s better to machine or weld first in a given case, and who plans the production? Is the AI going to coax a painter into staying for overtime again, since those large, variable parts aren’t being painted by any robot even today.
We also have food delivery robots, and the grocery stores aren’t empty yet. There are chatbots, yet people still want a human when problems arise because the bot can’t apply any context. Banks use automated decision-making for small loans, but apparently, the risk is too high since you don’t get a mortgage just by pressing a button.
Of course, working life changes and evolves, but I doubt the change is as fast in every field as people fantasize. In negotiations, people want to negotiate with humans, and quality has to be monitored by a human. Innovation also requires a human.
Related to the topic, I also asked the paid version of ChatGPT: Not a bad reflection.
AI’s impact on employment over the next 10 years is not binary (“increase” or “not increase”), but a macroeconomic balance between three forces:
automation vs. productivity growth vs. the emergence of new industries
Structural analysis below.
- Mechanism: how AI can increase unemployment
This happens if automation progresses faster than the economy can create new jobs.
High-risk job categories
Especially tasks involving:
repetitiveness
rule-based nature
text processing
standardized decisions
Examples:
customer service
entry-level analysis
translation work
administrative documentation
In these, AI can replace 30–70% of tasks without significant investment.
- Mechanism: how AI reduces unemployment
Historically, technologies have eventually increased employment because they:
lower production costs
increase demand
give birth to new industries
Historical examples:
electricity
the internet
industrial automation
All were feared → all created more jobs in the long run.
- Most likely scenario over a 10-year timeframe
Macroeconomic realism:
| Period | Impact |
|---|---|
| 0–3 years | local job loss |
| 3–7 years | structural change |
| 7–10 years | balance |
Meaning, in the short term, unemployment may rise in some sectors.
- Where unemployment is most likely to grow
Most likely sectors:
routine office work
basic analytics
bulk-level content production
junior-level roles
Important note:
AI replaces tasks, not entire professions
- Where employment grows
AI increases demand especially in:
technical design
monitoring and quality assurance
decision-making
creative work
physical jobs that cannot be automated
- The biggest risk is not unemployment, but this
Researchers consider the most likely impact to be:
growth in income inequality
Reasons:
productivity for high-skilled labor increases significantly
low-skilled work does not scale
- Economic consensus forecast
Summary of most labor market models:
on a global scale within 10 years
unemployment is unlikely to rise permanently
but the labor market will change drastically
Final analysis
Most likely realistic scenario:
AI increases temporary unemployment in some sectors, but does not lead to permanent mass unemployment globally over a 10-year period.
The biggest change is not the number of jobs but:
work structure
required skills
wage distribution
Condensed answer:
AI is unlikely to permanently increase global unemployment within 10 years — but it will change the labor market more significantly than any technology since the internet.