We are a network of students and professionals in Cambridge (UK) working on AI safety—the problem of ensuring that AI systems don’t cause catastrophic failures as they become increasingly powerful and complex.