More

    Amazon Steps in to Solve ‘AI-Penned Books’ Headache

    Published on:

    As generative AI continues to gain traction, Amazon has introduced the following policies: Books written by bots These have become a headache for e-commerce companies after they flooded the platform in large numbers.

    Generative AI’s ability to generate prose, poetry, lyrics, images, audio, and video in general is captivating the world, and people with little or no skills are turning to technology to generate content. Ta.

    “AI nonsense”

    But the same technology comes with its own set of complications that Amazon is in the process of fixing. According to an article, the “jack of all trades” is currently full of books written by bots. wired.

    This article does not say exactly how many AI-generated books are sold on the platform, but the platform has a wide range of books covering a variety of genres, with some The problem is low-quality titles. Earlier this year, author Caitlin Lynch called this “AI nonsense.” Post to X (Formerly Twitter).

    Even more concerning is that some books bear the names and likenesses of real human authors. For example, this summer, author and publishing industry professional Jane Friedman discovered five of her books incorrectly listed as hers on Amazon.

    One deceptive title is “How to quickly create and publish an e-book.” And the people behind these bots are determined in their unscrupulous activities. According to a report in Wired, there’s a “complete hustle culture on YouTube videos.”

    These encourage viewers to get rich easily by creating AI-generated content for Kindle, with titles like:

    “How to Create Undetectable AI Content for Amazon KDP (Guaranteed).”

    These modern money-making mechanisms make it difficult for readers looking for books to obtain quality reading material written by humans.

    Also read: ChatGPT’s real-time web browsing frees you from limited data in 2021

    Amazon’s policy

    Faced with this dilemma, Amazon has begun taking steps to address the problem in the past few weeks. The company currently has a cap on the number of books a single author can publish. Also, policy It would require authors to disclose whether their books were written by humans or generated by AI.

    “Amazon has an ethical obligation to disclose this information,” Friedman said.

    “Authors and publishers should already be disclosing this information, but if they aren’t, Amazon, along with all retailers and distributors, needs to mandate it.”

    Not doing so would amount to “creating distrust and confusion,” she added.

    Authors Guild CEO Mary Rasenberger agreed: “We will require platforms and publishers across the board to flag AI-generated material as such. “I’ve been advocating for legislation.”

    But there is another way

    However, there are other ways to address this issue, one of which is the use of AI detectors.

    Several AI detection startup People separately interviewed by Wired suggested Amazon could experiment with AI detection technology to close loopholes and make it easier to flag AI-generated material.

    Reality Defender, one of the AI ​​detection startups, says that while it started with a focus on deepfake images, it has now expanded to text detection as well, a service that other companies also offer. This could be the solution to Amazon’s problems.

    The CEO of another startup, GPTZero, has similar confidence.

    “We are definitely able to bring this technology to e-commerce platforms and are currently in discussions with various vendors to do so.”

    Winston AI founder John Renaud already has several publishers among his customers.

    Amazon also has an obligation to meet customer expectations for high-quality books written by humans, but there are also concerns about why the company hasn’t adopted tools to automatically flag books created by AI. . Amazon declined to comment for this story, but spokeswoman Ashley Vanicek issued a written statement.

    “Amazon is constantly evaluating new technologies and is committed to providing the best shopping, reading, and publishing experience possible for our authors and customers.”

    skepticism

    There is skepticism about the accuracy of AI detection tools and their results, which may be why Amazon is taking a cautious approach to AI detection tools.

    A paper published by researchers at the University of Maryland shows that the detector’s results can be flawed.

    “These detectors are unreliable in real-world scenarios,” they say. I have written.

    Another study conducted in July by researchers at Stanford University showed how the detector is biased against writers whose first language is not English.

    OpenAI, the maker of ChatGPT, had to shut down its AI classification following criticism over poor results.

    After a failed experiment with Turnitin’s AI detection program, Michael Coley of Vanderbilt University said, “We do not believe that AI detection software is an effective tool to use.”

    Related

    Leave a Reply

    Please enter your comment!
    Please enter your name here