Generative AI Student Tutorial for the Graduate School of Life Sciences University Medical Centre Utrecht Netherlands

We’ve made a substantial update to the student tutorial to help you better understand the current landscape of generative AI and its implications. Whether you’re just beginning or have already completed the tutorial, now is a good time to revisit it and explore the newly added material.

What’s New?

Updated and Expanded Section 2.0:

• New explainer on Artificial General Intelligence (AGI) and how it’s being benchmarked
• A full chapter clarifying the distinction between Large Language Models (LLMs) and broader Generative AI
• Introduction to diffusion models, including how they work and what they’re used for
• Expanded content on Prediction… Not Common Sense
• Expanded content on biases in training data and the social consequences of these biases
• A new chapter on AI sycophancy, covering how GenAI tools have been designed for agreeability and what that might mean when using these tools for feedback or assessment
• More detail on the Black Box problem and how neural networks obscure transparency
• New insights into deepfakes, including their social, political, and scientific implications
• An extended discussion of Environmental, Social, and Governance (ESG) issues related to GenAI development and deployment
• A new chapter on geopolitics and the global AI race, examining national strategies, ethical frameworks, and power imbalances
• Additional information of how GenAI and its impact on higher education,
• A new chapter about Looking Ahead and how we can be more proactive than reactive.  
• A brand new Test Your Knowledge quiz to reinforce and review your learning

Minor updates throughout:

• Improved clarity and updated terminology across Sections 3–5
• New chapter in Section 5.0 listing influential voices and organisations to follow in areas such as:

·      AI ethics

·      AI and sustainability

·      AI in biomedical and life sciences

·      Policy and regulation

·      Critical and responsible GenAI use

Introduction to the tutorial

As part of our commitment to innovation in research and education, the Graduate School of Life Sciences (GSLS) at University Medical Centre Utrecht has begun integrating the use of Large Language Models (LLMs) into our educational framework. The GSLS is renowned for its interdisciplinary approach, with 16 master’s programmes and a diverse community of 1,500 master’s students. One of our goals is to foster a collaborative and dynamic learning environment that prepares students for the challenges they may face in a modern life sciences career. In line with this mission, we recognise the growing influence of LLMs in academic research and education. To ensure that our graduates and staff can effectively navigate and utilise these tools, we have developed a tutorial designed to provide both the technical knowledge and ethical framework necessary for responsible use. This tutorial, which was launched in February 2024 alongside our new Master’s Student Guidelines, aims to assist students in the ethical and responsible integration of LLMs into their work while maintaining the highest standards of academic integrity and scientific rigour. The principal focus of the tutorial is fostering a human-centred approach to LLMs. At the GSLS, we view LLMs as tools to support human abilities, not as replacements for critical thinking or ethical judgment. Our goal is to ensure AI is integrated in a way that enhances, rather than diminishes, the essential human qualities of creativity, empathy, and critical reasoning in life sciences education and research. By embracing a thoughtful and balanced approach to AI, we reflect our core values of responsible research and open science, while empowering our academic community to innovate with integrity.

Attribution:

This tutorial is part of the GSLS GenAI resources developed by Christine Fox with the valuable contribution of Fleur Boelen and ChatGPT. We encourage educators and students to adapt and use this material in their learning and teaching processes, ensuring that any adaptations or shared versions are credited back to the original creators. We thank you in advance for respecting our efforts and contributions to the field of educational technology.

We would like to thank Karin van Es, Marie-Louise Goudeau, and Davitze Könning for their constructive feedback. Additionally, we thank Laura Huiscamp, Shirrinka Goubitz, Marit de Kort, Zoë de Wit, and Harold van Rijen for their reviewing, minor editing, and feedback. Finally, we extend our gratitude to Ruud Dielen for his publishing assistance.