DocuSign to use customer data for AI training, raising privacy concerns among users

DocuSign, the widely used electronic signature service, recently updated its FAQ to reveal that it may utilize customer data to train its proprietary AI models. The company clarified that it only uses data from customers who have expressly given consent, and the data is de-identified and anonymized before the training process begins.

The data usage applies specifically to certain products and programs, including DocuSign AI Extension for CLM, DocuSign AI Labs, and selected beta programs using AI. The company leverages both proprietary in-house AI models, which are based on customized open-source AI models, and third-party models such as Microsoft Azure OpenAI service.

The new disclosure indicates DocuSign's intent to enhance its AI program by incorporating user data, including the contracts they sign. This development comes in the wake of the launch of its Generative AI “Agreement Summarization” feature, designed to analyze extensive contracts and extract key points.

This strategy aligns with the recent actions of Reddit, WordPress, and Tumblr, which have been selling user data to train AI models from large corporations.