This pioneering collaboration aims to revolutionise the way datasets are used for training machine learning models. The focus is upon multilingual work essential in sectors such as medical, engineering, and safety. Inadequate or biased labelling can lead to AI 'hallucinations' – generating incorrect or prejudiced responses – which can have dire consequences in high-stakes industries.
