Foreword to Version 1.91

When the previous version of this Companion appeared, generative AI was something you had to visit a specific website for, those of for instance: ChatGPT, Gemini, Copilot or Claude. By now it is built into software you are already using, whether you asked for it or not. Think of WhatsApp or Tiktok.

In education, the conversation about it keeps returning to two questions: Do you learn less if you use AI? And does academic integrity hold up? There is no straightforward answer to either question, but we cannot avoid them either. That is why this version includes an additional chapter, Chapter 3, on the effect of generative AI on learning. It is a tightrope to walk: the same tool that can take over your thinking can also strengthen your learning.

Meanwhile, the world around it has not stood still. A handful of American tech companies are becoming ever more powerful, China is claiming a larger role, and the planet is warming up faster than expected. These are not footnotes to AI; they are forces of which your AI-use forms a part. In some places, the previous version of this Companion struck a more upbeat note than we now think responsible, and those passages have been revised.

In the final chapter you'll discover what you can do to contribute to an AI-aware future. Not as a checklist of personal obligations, but by organizing the conversation about AI at five levels: from how the technology works to how it is embedded in society and what your own agency could be. Once you know at which level a discussion is taking place, you can contribute to it. Further changes since version 1.9 include:

• New topics covered: anthropomorphism, the attribution of human qualities to systems that do not have them; agentic AI, where a language model reuses its output as input for a subsequent step; vibe-coding; and the AI bubble, about how overheated economic expectations can lead to a financial crisis.
• More on social media: the unwillingness of platform companies to curb addictive algorithms and to check false or inflammatory messages, with an example of how incitement is fuelled.
• Gemini Notebook has been added as an example of a tool that presents difficult material in several views side by side.
• The chapter on doing research with generative AI has been removed from the Companion. This means that programs that consider it undesirable for students to use AI in research can also use this book without reservation.

This Companion is neither a manual nor a warning. It is intended to help you understand enough about what happens under the hood, and about what happens around it, to choose your own course. What that course will be remains up to you.

Silvester Draaijer, Marianne Eggink and Charlotte Meijer, VU Centre for Teaching & Learning