The new GDPR regulation requires that any company must be able to prove that the personal data it holds are protected and, above all, unusable in case of theft. This has created a new need for automatic tools to identify and mask protected data, including in texts, in order to facilitate companies’ compliance with the legislation. The creation of such tools, that allow robust and versatile text processing to handle personal data, is still an important issue and requires the creation of specific semantic models for linguistic AI. This panel will outline the current landscape in the processing of personal data in texts, by providing the point of view of both researchers in Natural Language Processing (NLP) and actors of the private sector. It will also address the question of data governance related to personal data in texts. • What are the real needs of business when it comes to personal data processing for GDPR compliance? • What is the role of personal data governance for the creation of value? • How to create linguistic models for the processing of personal data? • What algorithms do we need for the efficient processing of personal data in texts?