From Hype to Tool
Few technologies have pushed their way into everyday work as fast as generative AI. Within two years a curiosity became a tool that drafts text, generates images and shortens routine tasks. And with adoption comes a new class of problem: no longer how do I get access? but how do I use this responsibly?
That is precisely where the workshop “Künstliche Intelligenz, KI verstehen & anwenden” comes in, which I deliver as part of the Sport Austria continuing-education programme. It is aimed at club members and officials — people who are not computer scientists but still have to make well-founded decisions.
Why Fundamentals Beat Prompt Tricks
Countless lists of “the 50 best prompts” circulate online. They are rarely useful, because they treat the symptom rather than the cause. Someone who understands how a language model works needs no recipe book.
The crucial point: a model like ChatGPT predicts, word by word, the most probable next piece of text. It does not “know” anything in the human sense and it looks nothing up. That explains both sides of the coin — the striking fluency as well as the tendency to produce convincing-sounding nonsense.
Once that has sunk in, the right questions follow automatically: where would the model have got this information? Is this a phrasing task or a factual question? AI is excellent at the former; the latter needs a source.
Text and Image: Two Very Different Tools
In the workshop we work with ChatGPT for text and Leonardo.ai for images. The use cases in a club setting are obvious: drafting circulars and newsletters, summarising minutes, structuring funding applications, generating social-media graphics, designing event announcements.
The format is deliberately interactive. You do not learn these tools by listening but by trying them — and above all by having something fail and understanding why.
Recognising Failure Modes — the Most Important Part
The most practically relevant section is the most uncomfortable one: AI makes things up. It invents legal provisions, studies, quotations and deadlines, and it does so in flawless, confident prose. This is called hallucination, and it is not a bug that will eventually be fixed — it follows directly from how the technology works.
For club work that means something concrete: an AI-generated text about association law, funding deadlines or tax questions is a draft, not advice. The rule of thumb from the workshop: the more verifiable a statement ought to be, the less you may trust it unchecked.
Data Protection Is Not a Formality
Anyone entering personal data into an AI tool is passing it to a service provider — often with servers outside the EU. Membership lists, athletes’ health data, job applications: none of that belongs in a chat window without consideration.
This is not an argument against AI but one for deliberate use. In the workshop we clarify which data is uncritical, where the relevant settings are, and when a local or European alternative is the better choice. Alongside that come the ethical questions: labelling AI-generated content, copyright in generated images, bias in training data.
What You Take Away
Not a certificate in “AI expertise”, but something more useful: a realistic picture of what these tools can do, where they fail, and how to apply them in your own environment without creating problems for yourself.
The next date in the Sport Austria calendar: 4 September 2026, 13:00–17:00 CET, online (delivered in German). No prior knowledge required.
We also run the same workshop as in-house training — tailored to your sector and your specific use cases. Details under AI Workshops & Training.