PhD Courses in Denmark

The emergence of capable generative AI models presents PhD researchers with both an opportunity and a challenge. The learning objectives, content, and course structure have been designed to promote critical thinking, technical competence, and responsible use of AI in PhD-level research. On one hand, tools such as large language models, code assistants, and multi-modal pipelines can dramatically accelerate research tasks that previously required significant manual effort — literature synthesis, exploratory data analysis, writing, and communication. On the other hand, uncritical adoption carries real risks: hallucinated references, eroded scientific voice, ethical breaches, and institutional policy violations. This course is designed to build genuine competence rather than surface familiarity. Students begin by grounding themselves in the ethical, institutional, and epistemological landscape of AI use in academia, and they progress through progressively more demanding hands-on projects that culminate in producing a short scientific paper from real research data. Each session pairs conceptual input with a concrete task, ensuring that the skills acquired transfer directly to students’ own PhD projects.

A student who has met the objectives of the course will be able to:

• Navigate the current landscape of generative AI tools
• Evaluate AI tools suitability for specific research tasks, taking into account performance, cost, privacy, and licensing considerations
• Interact with AI models programmatically through API calls and local deployments
• Design multi-step automated workflows tailored to research needs
• Integrate AI assistance into professional academic writing environments
• Apply AI tools to data analysis, visualization, and scientific communication using data from their own research
• Critically assess the quality, accuracy, and limitations of AI-generated text, code, figures, and interpretations
• Develop principled strategies for accepting, adapting, or rejecting AI suggestion
• Reflect on the broader ethical, epistemological, and societal implications of AI adoption in academic research
• Align their practice with DTU policy and evolving publisher standards

Students will develop a working understanding of the generative AI landscape, including the trade-offs between open-source and closed-source models, local and cloud-based deployment, and API-based versus interface-driven interaction. From this foundation, they will acquire hands-on experience designing and implementing multi-step AI workflows in Python, covering tasks such as batch document processing, cross document synthesis, automated translation, and structured output generation. A significant portion of the course is devoted to AI-assisted scientific writing. Students will learn to configure professional writing environments that combine LATEX, VSCode, and AI coding assistants, and will practise using these tools across all stages of manuscript preparation — from outlining and drafting to figure captioning, revision, and peer review. Critically, students will work with real data: either from their own PhD research or from publicly available energy-system datasets retrieved via the ENTSO-E Transparency Platform. This ensures that the skills developed are immediately applicable to their own work. Throughout, the course emphasizes responsible and transparent use of AI. Students will engage with the ethical, epistemological, and institutional dimensions of AI adoption in academia, and will develop criteria for evaluating and disclosing AI contributions in line with publisher and DTU standards.