I stumbled across this article regarding the Indian news agency ANI who, like similar news and media entities in other jurisdictions, also sued OpenAI for unauthorized use of their news content for training ChatGPT, but this time in a New Delhi Court. Fair dealing also exists in The Copyright Act of India although the enumerated categories in s.52 of that Act of activities that are considered fair dealing in India, are a bit different from Canada. Education, for example, which is a very broad category for fair dealing in Canada, is not a category that is explicitly listed in The Copyright Act in India. Thus, it would be interesting to see how this will play out in the New Delhi court as it seems that the categories of research and education would be strong arguments for OpenAI in Canada.
This particular piece also touches on an issue that I found particularly interesting – what happens if ingestion of this content is found to be copyright infringement and not fair dealing? Despite the fact that OpenAI has now stopped scrapping information from ANI’s sites, the news agency argues that the harm has been done already and the data is now “permanently stored in the memory of ChatGPT.” Data that has been used as input for such a complex training model, such as the one used by ChatGPT, cannot simply be removed, yet its benefit has already been extracted. It would be very interesting to see if there are other possible remedies, other than monetary damages, that might be available to a company whose copyright has been infringed in an AI ingestion context.
Indian Copyright Act: https://www.copyright.gov.in/Documents/Copyrightrules1957.pdf
Copyright & Social Media
Communications Law