Why should we be cautious when entering Data into Generative AI Tools?
Utilizing generative AI tools involves challenges regarding personal data protection and privacy. Many companies that create and offer these tools process the data contained in the content provided by users, sometimes reserving the right to share it with other entities (see, for example: OpenAI’s privacy policy).
Personal data – such as the names of students, social security numbers, index numbers, phone numbers, or email addresses – should not be entered into generative AI tools due to limited information about the data processing principles by service providers using artificial intelligence. This can be referenced in the statement from the Data Protection Officer of the University of Warsaw regarding the use of cloud storage:
"Due to emerging doubts regarding the storage of official documents processed in electronic form, I kindly remind you that the processing of all files containing official information, including personal data, is done solely using the tools provided by the employer."
Generative AI tools should not be used to process documents containing special categories of personal data, such as information on ethnic origin, union membership, or sexual orientation. Additionally, generative AI tools not provided by the employer should not be used to assess signed student papers.
