Artificial intelligence promises efficiency and quick solutions. Yet experts warn of a creeping process known in sociology as “de-skilling”: while machines grow ever smarter, humans lose essential skills.

The Anatomy of Competence Loss

The phenomenon is not new. GPS navigation already weakened our sense of direction, but generative AI is accelerating this process dramatically:

Less Creativity: When algorithms provide the first drafts, human, original ideation withers.
Loss of Knowledge: Those who no longer work out answers but merely consume them don’t anchor knowledge in long-term memory.
Blind AI Assistance: Critical questioning disappears. We trust machines blindly (“automation bias”), even when they hallucinate or error.

The Real-World Cost: The Lukas Scenario

Consider Lukas (28), project manager at a major Austrian tourism board. Task: Conceive a new “Sustainable Hiking Trail” with funding application.

Old way (expertise intact): Lukas studies maps for hours, calls landowners, checks GIS data for protected zones, researches local legends (“The Giant of the Valley”), connects them to local flora, understands political nuances between mayors and alpine cooperatives.

New way (de-skilled): Under time pressure, Lukas feeds AI: “Create a family-friendly, sustainable hiking trail concept for Austrian Alps, nature focus, for funding application.” AI delivers instantly: solid concept with stations like “barefoot path,” “insect hotel,” “wooden sound installation.” Professional language. Lukas adjusts place names, submits.

The Fatal Consequences

Loss of Regional Identity: LLMs train on averages of billions of data points. Result: generic “Alpine standardization.” The proposed barefoot path already exists in ten neighboring valleys. What’s missing: the old mine that could have been included, or the rare water orchid species unique to the site. Lukas no longer recognizes the region’s unique DNA.

Loss of Feasibility Competence: Because Lukas delegated research, he misses that the proposed route crosses a wildlife refuge (AI didn’t know) or a farmer’s land (who hasn’t granted rights for years). The project fails in implementation or approval is rejected.

Erosion of Stakeholder Management: Political diplomacy is crucial in Austria. Tone matters: formal government language vs. tourism marketing. Delegating to AI means losing the feel for what works with different audiences.

Three Countermeasures

1. Strengthen “Human-in-the-Loop”: AI is co-pilot, not autopilot. Education must shift from “write a text” to “have AI write it, then correct errors and weaknesses.” This requires MORE expertise, not less.

2. Deliberate Re-Skilling: Organizations should introduce “analog days”—work without AI assistance. A travel advisor must know where Vietnam is and how tariff systems work without algorithms. Only those who master basics can judge if AI performs well.

3. Foster Critical Creativity: Learn to ask questions AI can’t (yet) answer. Empathy, ethical judgment, and complex problem-solving in unforeseen situations remain human domains.

The Hard Truth

Ironically, cost-cutting often targets young talent, not long-term employees. But where do “old pros” come from if young people never gain expertise because AI replaced them? We’re creating a generational competence gap.

Conclusion

AI is powerful. But if we let it do our thinking, we shift from masters to spectators. The challenge isn’t making AI better—it’s staying good ourselves.

By Prof. DDr. Roman Egger / Smartvisions | On AI, Skills & the Future of Work in Tourism