coingeek.com
OpenAI and Microsoft (NASDAQ: MSFT) have found themselves in yet another legal battle concerning copyright infringement after being dragged to court by author and Hollywood reporter Julian Sancton.
According to the court filing, Sancton argues that the artificial intelligence (AI) developer relied on thousands of non-fiction books to train its large language model (LLM), breaching the intellectual property (IP) rights of authors.
Sancton, author of the New York Times best-seller Madhouse at the End of the Earth: Belgica’s Journey into the Dark Antarctic Night, said he spent a fortune on researching the facts for his book, earning “a bundle of exclusive rights” under the law. In the filing, Sancton argues that OpenAI and Microsoft have reproduced copyrighted works without compensating thousands of authors.
Using allegedly illegally obtained data, OpenAI trained its widely successful ChatGPT model with the IPs of thousands of non-fiction authors. After initially operating an open-source model in the first two iterations of its models, the filing argues that OpenAI opted for secrecy in the dataset construction and training methods for ChatGPT-4.
“Upon information and belief, OpenAI had another reason to keep its training data and development of GPT-3, GPT-3.5, and GPT-4 secret,” read the filing, stating that this is to keep rightsholders like the plaintiff and class action members in the dark about whether their works were being infringed and used to train OpenAI’s models.
The plaintiff alleged that OpenAI and its primary investor, Microsoft, raked in undue profits following ChatGPT’s skyrocketing adoption. With over 100 million weekly users and a premium ChatGPT Plus service costing $20 per month, OpenAI reportedly generates over $100 million in monthly revenue.
An integration of ChatGPT into Microsoft’s products like Bing has seen it record a spike in the number of new users, surpassing over 100 million active daily users for the first time. Researchers have opined that blockchain technology could solve AI’s copyright and data handling problems.
“While OpenAI and Microsoft refuse to pay non-fiction authors, their AI platform is worth a fortune,” read the filing. “The basis of the OpenAI platform is nothing less than the rampant theft of copyrighted works.”
OpenAI’s rough patch
OpenAI is undergoing a rough patch, accentuated by a leadership turmoil that rocked that company at the start of the week. The fiasco saw company CEO and founder Sam Altman
leave the firm to join Microsoft briefly and then returned as CEO of OpenAI.
The company has been juggling class action lawsuits, probes by data regulators, and several technical hitches in recent times as it grapples to stay ahead of its competition.