top of page

AI literacy: the tool is not the problem.

  • Aug 9
  • 4 min read

Updated: Aug 10

The implementation of a new AI-tool is only the first step. What decides its eventual benefit is the empowerment of its users.


Two members of staff use it for everything, including tasks it was never meant for. Two will not touch it, because they do not trust it. Practice management sees an invoice but no clear effect. And nobody on the team could explain, if asked, how this system actually arrives at its answers.


Both groups act for the same reason: no one ever showed them what this tool can do and where its possibilities end. That is not a technology problem. It is a competence gap.


The good news first: this gap can be closed, and without an IT project. What it takes has a name — AI literacy, the term used in the research. What that means exactly, and what your practice or your association office stands to gain from it, is the subject of this article.


AI literacy does not mean knowing how to operate AI


The common definition describes AI literacy as the abilities people need in order to live, learn and work effectively in an environment shaped by AI. That sounds abstract, but it means three very concrete things.


First, a basic technical understanding: knowing that these systems rest on statistical probability rather than on established knowledge — which is why they can state something factually wrong in entirely convincing language. Second, the legal framework: understanding which data may go into which tool and which may not. Third, an assessment of consequences: being able to judge what their use means for patients, for staff, for your own reputation and for society at large.


The single most important element is a distinction that is almost never drawn in daily practice: automation or augmentation. Automation means handing a step of the work over to AI entirely. Augmentation means dividing that step between human and machine so that each contributes its strengths: the AI produces the draft, the human checks it, corrects it and takes responsibility for it.


This distinction is not academic. In November 2025 the Stanford Digital Economy Lab analysed payroll data from the US provider ADP and found a relative decline in employment of around 16% among 22- to 25-year-olds in occupations heavily exposed to AI. The decisive part, however, is the second half of the finding: the decline occurred where AI replaces work. In occupations where AI complements the work, employment grew.


AI does not cost jobs. The way it is used does.


In Switzerland, too, this is no longer a question for the future


The International Monetary Fund estimates that around 40% of jobs worldwide are touched by AI. In advanced economies such as Switzerland the figure is roughly 60%, because more of the work here is knowledge and information work.


The Swiss eHealth Barometer 2026, conducted by gfs.bern, shows how the sector views this: 84% of the healthcare professionals surveyed can imagine automating routine tasks with AI in future. The report draws a remarkably clear conclusion — the benefit of AI depends directly on sufficient competence, and without targeted training its potential remains untapped.


The FMH, the Swiss Medical Association, takes a similar view. In its recommendations on the use of large language models it states that doctors should use such systems only where they are professionally able to recognise errors and judge the quality of the answers. No competence, no deployment. That is the short version.


There is also an obligation that many have not yet registered. Since 2 February 2025, Article 4 of the EU AI Act has required companies to ensure a sufficient level of AI literacy among their staff. The provision applies extraterritorially and can capture Swiss businesses as soon as there is an EU connection. Anyone who makes AI tools available to their team, or tolerates their use, must therefore be able to demonstrate that the team has been trained.


What your organisation gains


First, oversight and transparency. Once a team understands where the limits lie, the quiet use of private free tools stops. You find out what is genuinely going on in your organisation. That is the precondition for putting data protection on a sound footing at all.


Second, a benefit you can demonstrate. A trained team recognises for itself which tasks are worth using AI for and which are not. Instead of one tool that some overuse and others avoid, you end up with a small number of clearly defined applications with measurable time savings. For association offices this often lies in written work: consultation responses, member communications, translations into the national languages.


Third, a team that becomes more valuable rather than replaceable. Anyone who merely operates AI makes themselves dispensable. Anyone who directs it, checks it and takes responsibility for it increases their own contribution. On the evidence available today, it is precisely this stance that determines whether jobs disappear or are created.


And a fourth effect comes free of charge: the fear of the technology recedes. Research describes this «AI anxiety» as a widespread reaction to the shift — and building AI literacy as the most effective remedy against it.


The legal position, the evidence and the recommendations of the profession's own body all point in the same direction. It is not the next tool that determines the value of AI in your organisation, but the competence of the people operating it.


Ready for a conversation? Arrange an initial meeting: www.felber-advisory.ch/kontakt. Together we will look at where your team stands today, which competences are actually missing and what training would look like that fits your working day.


This article provides general information and does not replace legal advice in individual cases.


Sources

 
 
 

Comments

bottom of page