Microsoft Partners with HarperCollins for AI Training Causing Author Backlash
HarperCollins, a major player in the publishing world, faces pushback from authors after a proposed licensing deal surfaced, offering $2,500 per book to train artificial intelligence (AI) models. The agreement, which is voluntary and covers a three-year term, aims to use select nonfiction titles.
Bloomberg now identified Microsoft as the tech partner in this initiative, sparking discussions about the intersection of author rights and AI technology.
Microsoft reached a deal with HarperCollins that will allow the software company to use nonfiction titles from the book publisher to train its artificial intelligence models https://t.co/1Lb2rlHxZe
— Bloomberg (@business) November 19, 2024
Details of the Licensing Offer
According to Bloomberg, the deal with Microsoft only applies to nonfiction books only. The licensing terms outlined by HarperCollins emphasize a flat, non-negotiable payment per title, raising questions about fair compensation and the value of intellectual property.
Author Daniel Kibblesmith, known for Santa’s Husband, shared screenshots of the proposal for authors on social media, which included details about confidentiality and the tech partner’s identity. “I’d do it for an amount of money that wouldn’t require me to work anymore,” Kibblesmith noted, underscoring authors’ concerns about the potential undervaluation of their contributions.
Microsoft’s Role and Broader Context
Microsoft’s involvement in the deal points to its strategic efforts to enhance its AI technology by securing high-quality data for training models. Large Language Models (LLMs), the core of many AI advancements, need extensive text data to develop language capabilities.
Microsoft has not disclosed specifics about the AI models in development, but its collaboration with HarperCollins aligns with its broader initiatives in AI-driven content processing.
Author Backlash and Community Concerns
The reaction among authors has been strong, with many voicing concerns over creative control and compensation. Drew Broussard, another author, called on fellow writers to consult their agents and reject such offers preemptively. He pointed to a recent partial setback in the Authors Guild’s lawsuit against OpenAI, emphasizing that the courts may not always offer sufficient protection for content creators.
The lawsuit accuses OpenAI of using copyrighted material without proper authorization, highlighting ongoing issues in AI and copyright law.
Penguin Random House’s Contrasting Approach
While HarperCollins is exploring potential AI revenue streams, Penguin Random House (PRH) has opted for a stricter policy. In October 2024, PRH updated its copyright clauses to ban the use of its books for AI training without explicit permission.
This decision aligns with the European Union’s DSM Directive 2019/790, allowing content creators to opt out of text and data mining. PRH UK CEO Tom Weldon explained that the company prioritizes author rights while leaving room for controlled internal AI usage under specific conditions.
Diverging Paths Highlight Industry Divides
HarperCollins’ willingness to partner with Microsoft signals an attempt to explore new revenue models amid the rise of AI. In contrast, PRH’s firm policy against third-party AI usage shows a more cautious stance aimed at protecting author interests. These differing strategies underscore an industry grappling with the implications of AI on content ownership and revenue.
As AI continues to grow in significance, the publishing and media sectors will need to find ways to balance technological progress with fair compensation and content control. The varied responses illustrate the uncertainty that still surrounds how creative work will be managed in an era where AI’s role continues to expand.