When AI Automation Erodes the Skills It Was Meant to Support

AI systems can complete work accurately and still leave an organization less capable. A case from a Finnish financial services firm illustrates the risk: software that automated fixed asset accounting was scrapped after managers concluded that staff had lost the underlying expertise needed to do the work themselves.

The system reportedly worked as intended. But accountants told Professor Esko Penttinen of Aalto University School of Business that automation had stopped them from considering the principles behind their tasks. Executives became concerned that the lost competence threatened the firm’s viability, and the company retrained staff in fixed asset management accounting.

The case appears in The Vicious Circles of Skill Erosion: A Case Study Of Cognitive Automation , a study co-authored by Penttinen. It illustrates a broader concern about cognitive offloading: when people routinely delegate mental work to software, they may get less practice applying knowledge, checking results and making judgments. The effect can extend beyond individual workers to teams and organizations that no longer know how to operate without automated support.

That risk is especially relevant as generative AI moves beyond automating discrete, repetitive tasks. Its proponents argue that AI can take on parts of white-collar work and give employees more time for strategy. But systems that generate answers, interpret information or make recommendations can also take over steps that once helped workers build and maintain expertise.

AI is already being added to workplace tools and used in areas such as call-centre operations. In the UK, the government has backed a report projecting £46bn in savings from full-potential digitisation and appointed investor Matthew Clifford to make policy recommendations on AI. These examples point to growing pressure to deploy the technology widely, while raising questions about how its effects on skills will be measured.

The consequences described in the study are not inevitable. Automation can remove routine work without eliminating human understanding, but that depends on how jobs and systems are designed. If workers are expected to supervise AI, they need enough knowledge to spot errors, question recommendations and take over when a system fails. Training that focuses only on using a tool may not preserve the expertise needed to evaluate its output.

For technology leaders, the Finnish case suggests that productivity measures alone can miss a costly side effect. Organizations evaluating AI deployments may also need to track whether employees retain core skills, provide opportunities to practise them and clarify who is accountable for decisions. Keeping a human in the process is meaningful only if that person can understand and challenge what the system produces.

Concerns about technology weakening independent thought predate modern AI. In The Technological Society , published in 1954, French scholar Jacques Ellul argued that modern systems could erode freedom of thought. AI makes the question more immediate because it can participate in reasoning and judgment, not just physical or clerical tasks. The practical challenge is to capture efficiency gains without allowing the capability to think through the work to disappear.

Cited: https://www.telegraph.co.uk/business/2025/08/03/the-ai-revolution-is-here-to-make-you-stupid/

Your Data, Your Insights

Unlock the power of your data effortlessly. Update it continuously. Automatically.

Answers

Sign up NOW

info at aimaker.com

© 2024. All rights reserved. Terms and Conditions | Privacy Policy