Ethics, Data Privacy and Security, and FIPPA Considerations

New artificial intelligence (AI) tools and large language models (LLM) (e.g., ChatGPT, Google Bard, Bing Chat) interact in a conversational way and have many uses, but also present ethical challenges. Some key issues for educators related to AI-generated information are:

• Transparency as sources are not identified and accuracy cannot be confirmed
• Intellectual property issues due to unclear data ownership, authorship and citation information
• Low reliability as inaccurate or biased data may be sourced or fantasy responses provided
• Misinformation based on examples and data provided
• Lack of accountability since the source of data is unclear
• Privacy and security concerns as data could be misused or in breach of privacy requirements
• Perpetuating stereotypes, ageism, sexism and racism

Maintaining student privacy is essential. Check technology signup requirements and terms of use. To maintain privacy (and meetFIPPArequirements) ensure tools do not collect student data (e.g., phone number, email address, age). Tools that do not meet privacy requirements must be voluntary and alternatives provided.

Generative AI companies collect personal information from the time that a user visits the site to their completion of using their services. At minimum, account data includes enough information to associate the individual with their account to login (this is usually name and email address). Sometimes setting up accounts includes providing additional demographic data that is either optional or mandatory. For services that require payment, the payment information directly associates the individual based on how they pay with the account and associated content making it harder to anonymize or alias the individual.