Helmholtz Information & Data Science Academy (HIDA)

The growth of machine-learning conference submissions and rapid uptake of large language models have accelerated research, while leaving open the question of whether findings have become more reliable. A favourable random seed or contamination of a test set can be mistaken for a scientific advance. This lecture examines the resulting gap in trust, analyses common research practices that produce fragile conclusions, and discusses safeguards needed for machine-learning studies to make a durable scientific contribution.

AI coding tools now support code generation, error explanation, refactoring, implementation planning and debugging, but programmers lack a widely established practical framework for deciding how much control to retain. This lecture presents a framework spanning different levels of human direction and agent delegation. It distinguishes stages of planning, coding, testing, debugging, verification and documentation to help researchers, students and developers choose appropriate assistance for each task. The aim is to preserve understanding, reliability and control while using language models productively. The lecture also examines responsibilities and risks involving correctness, security, licensing, data protection and scientific accountability.