Practical guidance on the effective and responsible use of generative AI in academic work, research, teaching, and professional settings.
This resource covers key principles for working with large language models, the main risks to watch for, and curated further reading.
This page was created with assistance from ChatGPT and Claude, primarily for language refinement, organization, and structural improvements.
This resource is licensed under CC BY-NC-ND 4.0. Suggested citation: Krakowski, S. (2026). Field Notes on the Use of Generative AI in Academic Work, Studies, and Research. Mimeo. Stockholm School of Economics.
This resource is continuously updated based on research and practical experience. For tested prompts and custom instructions, see the Prompt Library resource. For structured innovation techniques, see the LLM Innovation Techniques resource.
This resource is a work in progress. Comments welcome.
Last updated: June 2026
Principles for Using Generative AI
Generative AI (GenAI) refers to a category of AI models - including large language models (LLMs) such as OpenAI's GPT series and Anthropic's Claude - that generate novel outputs based on patterns learned from training data. These models have demonstrated significant capabilities in understanding language, context, and intent, and are increasingly used across industries, including academic research, teaching, and professional knowledge work.
The technological, political, and legal context surrounding GenAI remains highly uncertain. Users should exercise caution, stay informed, and ensure their use complies with current regulations and institutional guidelines.
The following principles are based on research (notably Bowman, 2023) and practice (notably Mollick & Mollick, 2023a; 2023b), as well as the author's own experience.
Use the tools and experiment. 🧪 Learning to work with GenAI is an increasingly important skill. Try different tools and develop an intuition for how they work, where they help, and where they fall short. Start with tasks in your area of expertise, where you can evaluate the quality of the output.
Invest in your prompts. Minimum-effort prompts produce low-quality results. You will need to refine and iterate to get beyond mediocre output. This takes work. AI is unpredictable, and identical prompts can yield very different results across sessions. My best prompting advice is to approach prompting as a conversation and interact with the LLM as if it were human (knowing fully well it is not). I like to think of the models as aliens, emulating humans. 👽
Do not trust the output blindly. GenAI models are designed primarily to generate coherent and plausible language, not to ensure factual accuracy. Their output can appear authoritative and compelling while being entirely wrong. If the model gives you a number, a fact, or a reference, assume it may be incorrect unless you can verify it independently. This problem - commonly called "hallucination" - is not a bug but a fundamental feature of how these systems work.
Protect your privacy. A common misunderstanding is that these tools automatically train on, and/or disseminate, everything you submit. Information submitted to GenAI tools may be stored or used for model training, particularly on free tiers, but major providers such as ChatGPT now offer opt-out options even on free plans, and paid or enterprise tiers typically provide stronger guarantees, including no training on user data and automatic deletion of chat contents after a specified period. A useful rule of thumb when using free versions is to only share what you would be comfortable sharing in a public context. Check the current terms for whatever tool you use, as well as any applicable policies from your organization or institution.
Be thoughtful about applicability and bias. These tools are not appropriate for every task or context. Keep in mind that models are trained on historical data, and their output is shaped by the nature, representativeness, and processing of that data, along with design choices (such as guardrails and fine-tuning) made by the organizations behind each tool. The output can reflect and amplify existing biases.
Prompt Engineering
Prompt engineering refers to the practice of crafting effective inputs to get useful outputs from generative AI models. While the term can sound more technical than it is, the core idea is straightforward: the quality of what you get out depends heavily on what you put in.
Key principles
Be clear and specific. Provide context, state your goal, and define the desired format or tone.
Iterate. Treat prompting as a conversation with a human being (or a human-like alien 👽), not a one-shot query like Google or searching a document or a database. Refine based on what the model answers, such as "make it shorter", "focus on idea 2", or "stop talking like an AI".
Break complex tasks into steps. Smaller, sequential requests tend to produce better results than a single sprawling prompt.
Ask the model to explain its reasoning. "Chain-of-thought" prompting - guiding the model through intermediate steps - has been shown to improve reasoning quality (Wei et al., 2023), but requesting it through prompting seems to become less and less important as models become more independently capable.
In long sessions, models may lose track of earlier context. If responses degrade, clarify or start a fresh conversation.
For a practical collection of tested prompts and custom instructions, see the Prompt Library resource on this site.
Selected prompt engineering resources:
Anthropic (2023). Guide to Anthropic's prompt engineering resources. https://docs.anthropic.com/claude/docs/guide-to-anthropics-prompt-engineering-resources
Boonstra, L. (2024). Prompt engineering. Google Whitepaper. https://www.kaggle.com/whitepaper-prompt-engineering
Cohere AI (2022). Prompt engineering. https://docs.cohere.ai/docs/prompt-engineering
DAIR.AI (2023). Prompt engineering guide. GitHub. https://github.com/dair-ai/Prompt-Engineering-Guide
Hebenstreit, K. et al. (2023). An automatically discovered chain-of-thought prompt generalizes to novel models and datasets. arXiv. https://arxiv.org/abs/2305.02897
Hugging Face (2024). LLM prompting guide. https://huggingface.co/docs/transformers/main/en/tasks/prompting
Llama (2024). Prompting. Llama documentation. https://www.llama.com/docs/how-to-guides/prompting
Microsoft (2024). Promptbase. GitHub. https://github.com/microsoft/promptbase
Microsoft (2024). Prompts for education: Enhancing productivity & learning. GitHub. https://github.com/microsoft/prompts-for-edu
Mollick, E. (April 26, 2023). A guide to prompting AI (for what it is worth). One Useful Thing. https://www.oneusefulthing.org/p/a-guide-to-prompting-ai-for-what
Mollick, E. (November 1, 2023). Working with AI: Two paths to prompting. One Useful Thing. https://www.oneusefulthing.org/p/working-with-ai-two-paths-to-prompting
Mollick, E. & Mollick, L. (October 31, 2024). Stop writing all your AI prompts from scratch. Harvard Business Publishing Education. https://hbsp.harvard.edu/inspiring-minds/an-ai-prompting-template-for-teaching-tasks
OpenAI (2023). Prompt engineering. https://platform.openai.com/docs/guides/prompt-engineering
Schulhoff et al. (2024). The prompt report: A systematic survey of prompting techniques. https://trigaten.github.io/Prompt_Survey_Site
Shapiro, D. (2024). ChatGPT custom instructions. GitHub. https://github.com/daveshap/ChatGPT_Custom_Instructions
Wei, J. et al. (2023). Chain-of-thought prompting elicits reasoning in large language models. arXiv. https://arxiv.org/abs/2201.11903
Bibliography and Further Resources
General
APA Style Team (2025). Citing generative AI in APA Style. https://apastyle.apa.org/blog/cite-generative-ai-references
Athey, S. & Imbens, G. W. (2019). Machine learning methods that economists should know about. Annual Review of Economics, 11(1), 685-725.
Bowman, S. R. (2023). Eight things to know about large language models. arXiv. https://arxiv.org/abs/2304.00612
COPE (2024). Authorship and AI tools. Committee on Publication Ethics. https://publicationethics.org/cope-position-statements/ai-author
Goodfellow, I., Bengio, Y. & Courville, A. (2016). Deep learning. MIT Press. http://www.deeplearningbook.org
Lee, T. B. & Trott, S. (2023). Large language models, explained with a minimum of math and jargon. https://www.understandingai.org/p/large-language-models-explained-with
Microsoft (2023). Microsoft privacy statement. https://privacy.microsoft.com/en-US/privacystatement
MLA Style Center (2025). How do I cite generative AI in MLA style? https://style.mla.org/citing-generative-ai
Mollick, E. & Mollick, L. (2023a). Why all our classes suddenly became AI classes. Harvard Business Publishing Education. https://hbsp.harvard.edu/inspiring-minds/why-all-our-classes-suddenly-became-ai-classes
Mollick, E. & Mollick, L. (2023b). Student use cases for AI. Harvard Business Publishing Education. https://hbsp.harvard.edu/inspiring-minds/student-use-cases-for-ai
Mollick, E. (January 24, 2023a). The practical guide to using AI to do stuff. One Useful Thing. https://oneusefulthing.substack.com/p/the-practical-guide-to-using-ai-to
Mollick, E. (March 29, 2023b). How to use AI to do practical stuff: A new guide. One Useful Thing. https://oneusefulthing.substack.com/p/how-to-use-ai-to-do-practical-stuff
OpenAI (April 25, 2023). New ways to manage your data in ChatGPT. https://openai.com/blog/new-ways-to-manage-your-data-in-chatgpt
Wolfram, S. (2023). What is ChatGPT doing... and why does it work? Stephen Wolfram Writings. https://writings.stephenwolfram.com/2023/02/what-is-chatgpt-doing-and-why-does-it-work
Yang, J. et al. (2023). Harnessing the power of LLMs in practice: A survey on ChatGPT and beyond. arXiv. https://arxiv.org/abs/2304.13712
Zhao, W. X. et al. (2023). A survey of large language models. arXiv. https://arxiv.org/abs/2303.18223
Research ethics
Gatrell, C., Muzio, D., Post, C. & Wickert, C. (2024). Here, there and everywhere: On the responsible use of artificial intelligence (AI) in management research and the peer-review process. Journal of Management Studies, 61, 739-751.
Grimes, M., von Krogh, G., Feuerriegel, S., Rink, F. & Gruber, M. (2023). From scarcity to abundance: Scholars and scholarship in an age of generative artificial intelligence. Academy of Management Journal, 66, 1617-1624.
Hicks, M. T., Humphries, J. & Slater, J. (2024). ChatGPT is bullshit. Ethics and Information Technology, 26, 38.
Susarla, A., Gopal, R., Thatcher, J. B. & Sarker, S. (2023). The Janus effect of generative AI: Charting the path for responsible conduct of scholarly activities in information systems. Information Systems Research, 34(2), 399-408.
Videos
3Blue1Brown (2024). But what is a GPT? Visual intro to transformers. https://www.youtube.com/watch?v=wjZofJX0v4M
Karpathy, A. (2023). The busy person's intro to LLMs. https://www.youtube.com/watch?v=zjkBMFhNj_g
Karpathy, A. (2025). Deep dive into LLMs like ChatGPT. https://www.youtube.com/watch?v=7xTGNNLPyMI
Sanderson, G. (2024). Visualizing transformers and attention. https://www.youtube.com/watch?v=KJtZARuO3JY
Repositories
Calegario, F. (2024). Awesome generative AI. GitHub. https://github.com/filipecalegario/awesome-generative-ai
Google Cloud (2024). Generative AI. GitHub. https://github.com/GoogleCloudPlatform/generative-ai
Microsoft (2024). Generative AI for beginners. GitHub. https://github.com/microsoft/generative-ai-for-beginners
Reganti, A. N. (2024). Awesome generative AI guide. GitHub. https://github.com/aishwaryanr/awesome-generative-ai-guide
van Vaerenbergh, S. (2024). Awesome generative AI. GitHub. https://github.com/steven2358/awesome-generative-ai