Learners can access free Harvard artificial intelligence (AI) learning materials covering generative AI, prompt engineering and machine learning, although not every resource in a widely shared online collection is a free course.
The materials were highlighted by AI product professional Basia Kubicka in a LinkedIn post that brought together six Harvard-linked learning resources and four additional lectures and videos from the university's Computer Science 50 (CS50) programme.
A key resource is The Science and Implications of Generative AI, a Harvard Kennedy School course developed by professors Sharad Goel, Dan Levy and Teddy Svoronos. First taught in Spring 2024, the course examines how generative AI works, how it can be used and the opportunities and risks it presents to society.
Harvard provides a free archived version of the 2024 course for educators and learners. Its online materials include lecture videos, pre-class work and classroom activities, with sessions covering the introduction to generative AI, prompt engineering and applications beyond basic chatbots.
The prompt engineering session explores ways of guiding AI systems to produce more relevant responses. The session titled Beyond Chatbots examines system prompts, retrieval-augmented generation (RAG) and fine-tuning.
RAG allows an AI system to retrieve information from external sources to help inform its answers. System prompts, meanwhile, establish instructions that guide how an AI assistant responds. The wider Harvard course also examines the implications of generative AI, including risks, copyright, misinformation and the future of work.
The collection also points learners to Harvard resources on generative AI in teaching and learning, alongside introductory material intended to explain the technology. These should not all be treated as standalone courses, as some are resource collections and teaching guidance rather than full lecture programmes.
One item, Teaching with AI, is a Harvard Business Publishing seminar for educators, not a free course. The January 22, 2025 virtual seminar was listed at $495 per instructor and covered classroom exercises, teaching approaches and ways to help students use AI.
The four additional resources highlighted by Kubicka include the CS50x 2025 artificial intelligence lecture, a CS50 Extension lecture on AI and prompt engineering, a talk on how GPT-4 works and how to build applications with it, and a discussion titled LLMs and the End of Programming.
The CS50x lecture provides a technical introduction to topics including prompt engineering, decision trees, machine learning, deep learning, neural networks, large language models, transformer architecture and AI hallucinations. Its online materials include lecture videos, notes, slides and transcripts.
The GPT-4 discussion, published as a CS50 Tech Talk in May 2023, features Ted Benson and Sil Hamilton examining how the model works and how developers can build applications with it.
Kubicka recommended starting with the material on system prompts and RAG for learners with limited time, arguing that these areas help users move beyond basic chatbot interactions.
The collection offers several entry points for people seeking to understand AI, but learners should check individual resource pages for access conditions, course requirements and any fees before enrolling.
