Скачать или смотреть The Data Provenance Dilemma: Legal and Ethical Risks

3:00 - 3:45 PM, Skyline CD&E, West Tower, Floor 2

Moderator: Irving Wladawsky-Berger, Fellow, MIT Initiative on the Digital Economy
Shayne Longpre, PhD Student, MIT Media Lab
Robert Mahari, PhD Student in the Human Dynamics, MIT & JD candidate, Harvard Law School
Mike Mason, Chief AI Officer, Thoughtworks

Generative AI models are highly dependent on the quality and quantity of data used in their training. But, many of these models are being trained on vast, diverse, and inconsistently documented datasets that have been raising serious concerns about the legal and ethical risks involved. This is true not only for large language models trained on text, but also for models based on different kinds of inputs, including audio, images, video, and software. We're seeing an increasing number of lawsuits about the provenance of the data used in their training, and the recent strikes by the writers' and actors' guilds were partly due to making sure that their work isn't used to train AI models without their permission and remuneration. The panel will discuss the complexities of using large, diverse datasets in the training of generative AI models, the difficulties involved in understanding the provenance of the data being used, and the kinds of tools and standards needed for the responsible use of these powerful models.