VIIMSI: safe AI use and risk management (AI agents)

AI does not have to stop at answering questions. AI agents can carry out several work steps in a row: search for information, compare offers, tidy files and draft documents. You give the task, and the agent does the work.

But how do you decide what to share with it, which answer to trust, and when must a person confirm before the AI acts?

In this practical workshop you learn to recognise the risks of using AI and to make informed choices that protect your data and your organisation's. No prior cybersecurity knowledge is needed.

We look together at: which data is public, internal or sensitive, and what may be given to AI at all; how to reduce the amount of data shared and make a safer copy of a work file; which account, storage and connected-app settings are worth reviewing; how to check the key facts in an AI answer and spot misleading content; how to limit AI's access and permissions, and when a person must approve a decision. We solve examples from everyday work. At the end you take away a short checklist for assessing your next AI use case: what data, what access, what could go wrong, and where human approval is needed.

For people who use AI in work tasks or want to start, and want to do it more safely. Free.

Trainer: Krister Viirsaar, an experienced software developer and trainer with over 10 years in IT, who teaches practical skills, understanding of how technology works, and conscious, responsible use of AI.

Part of Arunädal and the Eesti.ai initiative 'The most AI-savvy nation'.

Wednesday 18 November, 17:00–20:00, Viimsi Library, Randvere tee 9, Haabneeme.