Dr. John Licato (CSE, assistant professor) was interviewed by Computerworld, an online magazine, about concerns raised over OpenAI’s submission to the British House of Lords. A key statement under fire seems to indicate their position is that GenAI is valuable enough to society to warrant special provisions in copyright law.
“Because copyright today covers virtually every sort of human expression — including blog posts, photographs, forum posts, scraps of software code, and government documents — it would be impossible to train today’s leading AI models without using copyrighted materials,” said OpenAI.
A range of lawsuits have been filed in multiple countries over what many assert is OpenAI’s unlawful use of web-scraping, the automation of copying data from websites. OpenAI states it has made efforts to allow creators to opt out of letting the models train on their work, but that seems to start with the assumption that copyrighted material was and remains fair use for them. This raises the fair question: what are some alternative methods of training AI without copyrighted material?
To find out about Dr. John Licato’s view on this topic please see the full article on Computerworld, click here.