By Peter Morici
Giving the writers what they want would stifle competition
Most of the literature that chatbots will read is in the public domain and on the internet.
Fiction writers are in an uproar that artificial intelligence will intensify competition. They want large-language models to either stop training on their literature or to be paid royalties.
Like the Luddites weavers in the 18th century, who broke into factories to smash textile looms, these writers want special status to protect their employment from technological progress.
Chatbots built on generative AI technology such as OpenAI’s ChatGPT and Alphabet’s (GOOGL) Bard can train by reading whatever is on the web — including newspapers, industrial diagrams, open-source software, paintings, textbooks, the classics and modern fiction.
Those programs — with errors humans must correct — can write accounts of historical events and news stories if fed breaking information like press releases or economic reports. They can also write knockoffs of contemporary fiction.
An Authors Guild open letter to the CEOs of OpenAI, Microsoft, Alphabet, Meta, Stability AI and IBM — signed by novelists Nora Roberts, James Patterson, Margaret Atwood and thousands of other writers — asks that software creators obtain permission and pay fees to train large language models with their work, even when copyright laws are not violated.
Importantly, most of the literature that chatbots will read is in the public domain and on the internet. To the extent that sites post illicit copies, writers’ complaints and copyright suits should be targeted at illegal postings and the owners of those sites.
Students prepare for writing careers — be it for the newspapers or penning novels — by reading both classics and contemporary authors. They don’t have to pay Atwood or Mark Twain’s heirs to access their books at the library or pay royalties as they publish throughout their careers.
The same goes for many professions — economists read Adam Smith, David Ricardo and contemporary journals available on the internet.
Regarding fiction, when authors or programs appropriate whole settings and characters to generate knockoff work, writers have copyright protection and the courts to obtain redress.
When chatbots generate work that is qualitatively distinct, those programs are no different than living writers or artists who elaborate on what preceded them. I doubt any successful contemporary author writer has not read Ellery Queen, Agatha Christie or Arthur Conan Doyle.
The problem for the novelist, artist, industrial designer, economists, coders and most anyone earning a living manipulating symbols, words and numbers or putting images on canvases and screens is that AI programs are becoming much more efficient than we are.
Read: How AI will make your job more satisfying and productive
We learn to apply those programs to the repetitive elements of our work and focus more on creative aspects. But in the process, competition reduces our numbers in the workforce and some of us must move on to other lines of work. That’s how new technology frees up labor, increases productivity and generates economic growth.
For example, Grzegorz Rutkowski has studied the great masters, such as Caravaggio, Rembrandt and Vermeer, as well as contemporary artists, to mimic their techniques and become an in-demand illustrator of beasts and landscapes for the videogame industry.
Now AI programs have mimicked him along with other popular artists, and he has joined a class-action suit against several of the companies that developed these systems. But it’s hard to see how Dall-E 2 or Stable Diffusion, created by OpenAI and Stability AI, respectively, are any different than an aspiring art student or how Rutkowski embroidered on the masters’ techniques to gain fame.
It’s one thing for a chatbot to so appropriate language, characters and settings, and another to create something substantially new in a similar style.
Prohibiting chatbots from reading work or requiring royalties on what those create would be like requiring aspiring fiction writers at Columbia University to either not read the work or be taxed throughout their careers for anything that appears inspired by an author’s style. The price of success is that others will emulate — even critics will appropriate it to improve on it.
The Authors Guild letter asserts “generative AI threatens to damage our profession by flooding the market with mediocre, machine-written books, stories, and journalism based on our work.”
That’s the rub — much of what prominent authors do is somewhat formulaic. Successful authors will continue to succeed by creating truly original work that strikes a chord with contemporary culture. Great writers carry notebooks or electronic devices to record what they see, or have remarkable memories, to create something new and vital. AI can differentiate, but I doubt it can inspire.
Peter Morici is an economist and emeritus business professor at the University of Maryland, and a national columnist.
More: Halloween won’t be the same because of the actors strike. Here’s why.
Plus: The Hollywood actors strike hits 100 days. Why hasn’t a deal been reached and what’s next?
-Peter Morici
This content was created by MarketWatch, which is operated by Dow Jones & Co. MarketWatch is published independently from Dow Jones Newswires and The Wall Street Journal.
(END) Dow Jones Newswires
10-28-23 1326ET
Copyright (c) 2023 Dow Jones & Company, Inc.
Source link
credite