“Publish or perish” is a pressure faced by all academic researchers. For African scientists, additional challenges arise: less access to tools, fewer publication avenues and limited visibility. At Agent Majeur, we partnered with the Swedish University of Agricultural Sciences and the University of Rwanda to deliver a cutting-edge “Writing with AI” workshop for 25 Rwandan agricultural scientists. The goal was ambitious: empower them with advanced AI training, thereby boosting scientific communication and safeguarding academic integrity.
Increasing competitiveness
Already under-represented in leading international journals, African scientists now face the prospect of falling further behind as AI becomes integrated into every aspect of research.
For Rwandan researchers, mastering generative AI tools is essential to gain international visibility. At Agent Majeur, we strive to give everyone equal opportunities. This is why the workshop was offered at an accessible price, as part of our Corporate Social Responsibility (CSR) program.
Learning advanced AI techniques
The workshop began by demystifying large language models (LLMs), these generative AI tools, and explaining how they are trained.
The program then focused on the art of prompting, moving from simple instructions to Few-Shot examples and advanced techniques like Prompt Chaining and Tree-of-thought to elicit precise, high-quality responses.
Mastering the double-edged sword
However, our trainer showed both sides of the coin. These tools, while fast, affordable, on-demand, multilingual, and multimodal (working with text, images, code and math), are not without risk. They can reflect biases from their training data, pose privacy and plagiarism concerns, and even generate sycophantic untruths—a phenomenon known as ‘hallucination’. Not to mention the MIT-termed ‘cognitive debt’—equivalent to relying on a walking stick when you don’t really need one, unduly weakening your legs. Each participant discovered practical ways to mitigate these risks.
25
Researchers
1
Day of intensive training
10+
AI tools mastered
Go forth, prepared
The researchers left not just with theory, but a toolkit. They explored specialized AI research assistants like Elicit and Consensus for literature reviews, and SciSpace for navigating complex papers. They also studied timeless methods for scientific English, to both improve their prompts and polish the outputs. A key insight came from combining the strengths of different tools: e.g. using LLMs with systems like LaTeX for precision in equations, tables and references or Python for creating flawless visualizations.
By the end of the session, all researchers were equipped to safely use AI with confidence, enhancing both the quality and international standing of their work. This training testifies to how Agent Majeur bridges talented local researchers with global scientific standards. And this is only the beginning. More AI-focused collaborations are already in the works.
These projets might interest you
