anonymization

This institutional and research caution is partly due to the fact that oral data is harder to anonymise than written data, and partly in recognition of the fact that in the future anonymisation will become increasingly harder as the application of AI to speech recognition continues to improve.