2026 CURIOSOIL Training – Prompting with Purpose: AI Tools for Science Communication

The CURIOSOIL project began 2026 by turning its attention to a timely and fast-evolving topic: how artificial intelligence can support communication, collaboration, and content creation in soil education work.

Led by Dr. Mihaela Bozukova from the European Science Communication Institute (ESCI), the session brought together CoP members for an interactive workshop that explored both the creative possibilities and the responsibilities that come with using AI tools. Rather than focusing on technology for its own sake, the discussion centered on how AI can support the mission of CURIOSOIL — strengthening soil literacy and making communication more accessible, effective, and engaging.

You can find:

• The handout
• The powerpoint presentation
• The recording below

From curiosity to practice

The session opened with a simple but telling observation: while AI tools are widely discussed, not all teams integrate them into their daily work. For some participants, AI is already part of weekly routines — used for drafting posts, summarising documents, generating images, translating materials, or brainstorming ideas. For others, it remains a new and somewhat uncertain territory.

Mihaela framed AI not as a replacement for human creativity, but as a collaborator. She presented five broad areas where AI tools can offer support: content creation, simplification of complex information, acceleration of repetitive tasks, expansion of ideas, and reflection on existing work.

This framing resonated strongly with the Community of Practice, whose members often juggle multiple roles — from coordination and facilitation to writing and dissemination.

Learning to speak AI: the RAFT framework

A central part of the workshop focused on improving the quality of interaction with AI systems. Mihaela introduced the RAFT framework — Role, Assignment, Format, Tone — as a practical way to structure prompts and guide AI tools more effectively.

Participants experimented with different prompting strategies, including zero-shot, one-shot, and few-shot approaches. Techniques such as iterative prompting and “chain of thought” reasoning were explored as ways to refine outputs collaboratively rather than expecting perfect results in a single attempt.

The session included a hands-on exercise in which participants applied these techniques to tasks from their own to-do lists. One example shared was the challenge of drafting a LinkedIn post — finding the right balance between clarity, structure, and visual elements without overusing formatting or emojis. Through experimentation, participants observed how small changes in prompt structure significantly influenced results.

The overall takeaway was clear: the effectiveness of AI tools depends largely on the clarity of human intention and guidance.

Responsibility and transparency

Beyond productivity, the workshop also addressed a crucial dimension: responsible AI use.

Mihaela outlined a practical workflow for safe engagement with generative AI tools, including:

• Preparing input data carefully
• Choosing GDPR-compliant platforms
• Fact-checking outputs before publication
• Being transparent when AI has supported content creation

Participants discussed different platforms — including ChatGPT, Gemini, Claude, and Le Chat — and reflected on concerns such as data privacy, environmental resource consumption, and the long-term implications of AI-generated content contributing to future model training.

Particularly relevant for NGOs and educational initiatives, the conversation emphasised the importance of informed tool selection and mindful usage. AI can accelerate work, but it does not replace critical thinking, verification, or ethical responsibility.

Building shared capacity

As the session concluded, several practical next steps were agreed. Mihaela will share the slides and a handout with participants to support further learning and internal dissemination. Participants were encouraged to begin building their own “prompt libraries” — saving effective prompts for recurring tasks and refining them over time. They were also reminded to activate opt-out settings in AI tools where possible to prevent input data from being used in future training.

The discussion also sparked interest in exploring fact-checking tools and deepening collective understanding of safe AI practices.

More broadly, the meeting reflected something essential about the CURIOSOIL Community of Practice: its willingness to explore new tools while grounding innovation in shared values. As soil education evolves in digital and hybrid spaces, AI may become one of many instruments that support clearer communication and wider engagement — provided it is used thoughtfully.

By starting the year with reflection on AI, the Community of Practice signaled not just technological curiosity, but a commitment to learning together, adapting responsibly, and strengthening the ways soil knowledge is shared across Europe.

Final Thoughts

The next meeting is scheduled for the 16th of March 2026, register now for the 2026 Spring meeting of the CURIOSOIL Community of Practice

Be sure to follow our  CURIOSOIL LinkedIn page and subscribe to our LinkedIn CURIOSOIL Newsletter.