A startup that uses artificial intelligence (AI) to create books has raised $37 million in the latest sign of how machine learning is transforming the publishing industry.
Inkitt’s self-titled app allows users to self-publish stories. Utilizing AI and data science, it identifies stories for refinement and distributes and sells them on another app, Galatea. The recent funding round for the company shows a growing interest in using AI for publishing. But while the technology brings opportunities for writers and readers, it also triggers concerns.
“For writers, the impact is profound but also dual-sided,” Zachary Weiner, CEO of Emerging Insider Communications, which focuses on the publishing industry, told PYMNTS in an interview. “The arsenal of tools at their disposal has expanded exponentially, empowering them to unlock creativity and concept with AI-generated prompts and ideas. They don’t have the constraints of conventional brainstorming as they can use AI to spark story inspiration and weave together disparate concepts in a more complex way. Moreover, AI offers an easier capability for proofreading and revising to streamline their workflow and focus on their core work.”
However, Weiner cautioned that AI is also bringing challenges to publishing. “AI is encroaching upon all writing and content jobs in some fashion, including editors,” he said. “As organizations strive to cut costs and maximize efficiency, the displacement of human writers by AI is already active and continues to be inevitable.”
The landscape of generative AI is rapidly expanding, propelled mainly by pioneering efforts from companies like OpenAI and its contemporaries.
Inkitt was founded in 2013, the company said in an August 2023 press release. It provides a platform for authors to share their work, engage with readers, and grow their audience. In 2016, the company achieved a milestone by selling “Bright Star” by Erin Swan to Tor Books, part of Macmillan Publishers. The company said this was the first book predicted to be a bestseller by AI, showcasing Inkitt’s approach to using data analytics in publishing decisions.
For publishers like Inkitt, adopting AI promises to streamline operations and reduce the reliance on personnel, Weiner said. Tasks that once required a legion of editors, proofreaders and layout designers can now be accomplished with a fraction of the workforce, leading to significant cost savings for publishers.
“And in the digital age and following a slew of industry failures and difficulty in monetization models — this is incredibly compelling; however, outsourcing crucial tasks to AI introduces a host of challenges, ranging from concerns about accuracy and reliability to the preservation of editorial integrity,” he said.
An influx of AI-generated books has inundated self-publishing platforms. The platforms, which account for a substantial portion of book sales, see around 34 percent of all eBooks being self-published, according to WordsRated, a publishing industry research firm. The proliferation of content on these platforms coincides with the widespread availability of AI applications powered by large language models (LLMs).
Sol Nasisi, founder of Booksie, a site for writers to self-publish that includes AI tools, said in an interview with PYMNTS that publishers must adapt to the new technology.
“Incorporating AI into a business process, especially in publishing and writing, will become easier, and companies that don’t figure out how to make their products better, faster, or cheaper using AI will suffer from the competition,” he said. “In addition, LLMs and generative AI will continue to advance and become more powerful.”
Nasisi said that he is concerned AI will train its models using the work of the site’s authors without their knowledge and consent.
“This needs to change as the technology matures so that content creators can choose how their content is used and are fairly compensated,” he said.
While AI excels in certain areas, such as content optimization and data analysis, its capabilities remain limited in others, particularly regarding nuanced tasks, Weiner said.
“AI-driven content also raises some heavy questions about authenticity and originality,” he said. “As publishers increasingly rely on AI-generated content to meet demands, there’s a risk of diluting the unique voice and perspective that human writers bring to the table.”
Not everyone thinks AI is something for writers or readers to fear. Using AI to aid in writing is akin to deploying a grammar checker or having an editor take a hands-on approach to manuscript changes, John Misak, an English professor at the New York Institute of Technology, told PYMNTS in an interview.
“If the quality of the book is high and the idea unique, I consider the work an addition to literature,” he said. “In addition, many ‘bad’ writers created great works with the help of a hands-on editor. AI can help make ‘good’ writing more equitable when used properly.”
The main danger of AI comes from scams that can be easily generated, such as copying other authors’ work, creating fake news, and various other deceptive schemes that AI can quickly produce, Misak said.
According to Misak, where readers place blame if something goes wrong is the main factor shaping how AI impacts publishing.
“Will readers understand AI content comes from a few bad players, or will they think publishers en masse are guilty of fabrication?” he said. “If they blame the few, this can serve as a protection via reviews. If they blame publishers and suspect all writing as AI-generated, much like many professors might with student work, then the problem spreads to the whole industry.”