AI Is Being Trained Using Human Creative Output. Is That Fair?
Eleanor M. Lackman will be speaking on the panel, "AI Is Being Trained Using Human Creative Output. Is That Fair?" on May 12, 2020 for the KnowIt Conference.
About the conference...
KnowIt is a groundbreaking new annual event that brings together the community of creators and protectors of intellectual property and other innovative outputs for an unprecedented discourse and collaboration. With dramatic shifts in both consumer and business technologies, as well as social trends and business models, now is the time for KnowIt.
Attendees at KnowIt will include lawyers (in-house and outside counsel), academics, inventors, authors, brand creators, cybersecurity and privacy professionals, leaders in government, regulators, startups, tech companies, investors, non-profits, media, Wall Street analysts and more.
Join us virtually on May 11-13, 2020, as we define intellectual output in a digital world and usher in the Fourth Industrial Revolution! *Due to the continuing impact of COVID-19, and our desire to unite the ENTIRE community of innovators and IP professionals, we will provide full access to KnowIt 2020's live stream event for FREE.
"AI Is Being Trained Using Human Creative Output. Is That Fair?"
Tuesday, May 12, 2020
4:35 PM - 5:20 PM
Trademark & Copyright
AI training, which is a fundamental part of any AI project, has been in the news of late. For example, a Google AI was trained to speak more like a human by being fed human-created novels. A Sony/Flow Machines AI was trained to create music by being fed human-created songs. And an IBM AI was trained to recognize faces by being fed human-created photographs taken from Flickr. Even assuming that AI tools need to be trained this way for now (until it can emulate the more top down reasoning that more closely resembles the way that humans approach problems), there are still questions about what is, and what is not, fair. Should companies that use copyrighted material to train AI be required to ask permission first, and/or to compensate the underlying creators? Or is the AI training activity a fair use? If it’s not fair use, how do parties that want to “consume” creative works in this fashion go about seeking consent from hundreds or even millions of creators? What about larger societal issues? Some have called for changes in law and policy to promote social equity in the AI age, such as universal basic income and “robot taxes.” These are all questions that still need to be answered. Don’t miss this fascinating discussion about copyright’s place in the Fourth Industrial Revolution.
Lauryn Guttenplan, Deputy General Counsel, Smithsonian Institution
Eleanor M. Lackman, Partner, Mitchell Silberberg & Knupp LLP
Nick Aries, Partner, Bird & Bird LLP