AI Data Permission: Do You Have the Rights to the Underlying Data & Information?*

AI Data Permission: Do You Have the Rights to the Underlying Data & Information?*

View Session Speakers

From a legal risk perspective, creators and users of artificial intelligence (AI) and machine learning (ML) innovations need to be careful with acquiring information, whether it be for the purpose of selecting the training data used to train these systems, or by acquiring or scraping copyrighted data or images without authorization to create a corpus of data that will be served to others for profit— or otherwise.

This panel will tackle the fundamental question that so many coders and developers would rather ignore: Do you have the legal right to be using the data and information— whatever that data and information may be— within the systems being created?

The panel will be split into roughly three segments. First, there will be a discussion of the technologies used and ways in which data and information are being acquired, which sets the foundation for the legal issues to be discussed.

Second, the panel will confront the issue of training data developers are using to teach AI and ML innovations. Do they have permission to use the data at all? And if they have the legal right to access the data do they have the right to derive economic gain? These can be particularly concerning questions, especially when using data from customers, but also when using purchased data. Consideration must be given to the scope of usage rights and limitations given privacy laws, which can be vastly different from state to state and country to country.

Finally, the panel will tackle the question of using copyrighted data and images. Under what circumstances is it acceptable to use information and images? When, if ever, is the use of copyrighted information and images a fair use?

Materials

Machine Learning Model: The Legal Need for Editability (Part I and Part II)