UCL and AIMS release free university teaching toolkit for AI foundations

The classroom-ready resource adds more than 50 hours of lesson plans and activities to Google DeepMind’s AI Research Foundations curriculum, combining technical AI teaching with ethics and wider social questions

UCL and AIMS have launched a free AI Research Foundations Toolkit with Google.org funding, adding 50+ hours of university teaching materials to Google DeepMind’s AI curriculum

University College London and the African Institute for Mathematical Sciences have launched a free teaching toolkit designed to help universities turn Google DeepMind’s online AI Research Foundations curriculum into a blended classroom program.

Funded by Google.org, the AI Research Foundations Toolkit contains more than 50 hours of teaching materials, lesson plans and classroom activities. It was co-designed with African educators and can be adapted by universities to different curricula and teaching environments.

Google DeepMind’s AI Research Foundations already provides an online learning pathway for students and community learners. The new resource is aimed at educators who want to build structured teaching around that material rather than simply directing students to complete the course independently.

The underlying curriculum was developed by UCL in collaboration with Google DeepMind. It is intended for university students and community learners who are proficient in Python and studying computer science, mathematics, physics or other technical subjects.

From language models to model fine-tuning

The original learning pathway takes students through the technologies behind modern language models such as Google Gemini.

Individual courses cover the fundamentals of language models, preparing and representing text data, neural network training, transformer architecture, model fine-tuning and the role of computing hardware in accelerating model development.

Students can also complete practical work including building and training a small language model.

The pathway finishes with a capstone course in which learners apply technical knowledge, ethical awareness and problem-solving skills to the development of a model for a real-world use case.

The new teaching materials are designed to make those online components easier to incorporate into university classes.

Rather than treating AI purely as a computing subject, the toolkit also brings together machine learning with ethics, sociology and philosophy. UCL Knowledge Lab describes the materials as modular, allowing educators to select and adapt individual components rather than adopting a fixed curriculum.

AI ethics built into technical teaching

The ethical component runs alongside the technical content rather than being positioned as a separate course. The toolkit is designed to support teaching around both how AI systems work and the responsibilities involved in developing them, including questions around the social effects of AI technologies.

That approach reflects the collaboration behind the resource. UCL and AIMS developed the toolkit with educators in Africa, while Google.org provided funding. The resource is available as an open educational toolkit for educators globally.

Google DeepMind’s underlying AI Research Foundations learning path currently consists of eight activities and is managed through Google Cloud’s learning platform. Its courses include a six-hour introduction to building a small language model, modules on data representation and transformer architecture, an eight-hour fine-tuning course and a 15-hour capstone project.

The AI Research Foundations Toolkit is available free for universities and educators to use and adapt alongside the existing online curriculum.