CIGI Papers No. 290 — March 2024 - Data Disquiet

Data Disquiet: Concerns about the Governance of Data for Generative AI 1 companies that control the use and reuse of much of is comprised of data sets created, collected or the world’s data (Staff in the Bureau of Competition acquired by the model developers. [...] Data Disquiet: Concerns about the Governance of Data for Generative AI 3 Box 2: Key Words Data provenance: Entails providing information on the origin of the data underlying a model and any changes or modifications the data set has undergone, and details supporting the confidence or validity of the data. [...] In on AI) that it would ask the director of the US a companion document to the act, the Canadian Copyright Office to issue recommendations to the government says that firms must document the president on potential executive actions relating data and algorithms these firms utilize and assess to copyright and AI. [...] According to the state of Meanwhile, Canada’s Directive on Automated the art at the time of entry into force of Decision-Making governs a wide range this Regulation, the cumulative amount of AI systems procured by the Canadian of compute used for the training of the government. [...] Recital 44 of the law notes that “datasets for more aware of enforcement problems and training, validation and testing, including the labels, should be relevant, governance gaps, leading to disquiet over how sufficiently representative, and to the best extent possible free of errors and complete in view of the intended purpose of the system.” It should data is governed.