Category: AI

AI-Generated Books

Is there anything that AI can’t do (badly)? I am a bit fascinated and alarmed to notice that AI-generated books are readily found online, available for purchase. If you’re not sure what a book written by AI looks like, here’s what I’ve learned so far:

Amazon, Of Course

Where will you find AI-generated books? Your first hunch would be correct: Amazon. They can be bought as dead-tree editions or as downloads for your e-reader, just like regular books. But they don’t disclose that they were made by AI. Here are some example listings for your consideration:

I’ve checked a few other online bookstores and not found signs of these types of books (yet). I suspect Amazon is the current marketplace for AI-generated books, because they make it so easy for anyone to self-publish and sell written works. And while it appears that authors must disclose AI-generated content to Amazon, I cannot find any mention that consumers must be informed. That means: It’s on you to distinguish books written by humans vs. those written by bots!

Why is This Attention-worthy?

At this time, I don’t want to levy any harsh criticisms on AI. Artifical intelligence can be used for good: to synopsize lengthy works, debug complicated code and even detect cancer earlier. But regarding these books being sold through Amazon, I have misgivings.

AI has a history of making odd mistakes. From recommending glue in a pizza recipe to misspelling “strawberry“, it’s clear that AI has some unexpected limits and flaws. Those who are using AI as a tool should do so mindfully, and be prepared for the occasional oddity.

I don’t presume the “authors” of AI-generated books are being careful or mindful. Rather, this looks like the low-effort work of lazy people looking to make a quick and easy buck. Lazy people typically don’t check their work. They don’t care about the harm that they could cause. Some examples of the damage they can cause are:

A Spotter’s Guide to AI-Generated Books

I imagine I’ll be revisiting and revising this, as these books evolve, but for now, here’s what tips me off to a book that may have been written by an AI:

AI-Generated Books
AI-generated headshot, created at Canva.com for free in under 30 seconds
  • title of book uses the word “Bible” to suggest it contains everything you’ll every need to know
  • volume claims to be “5 books in 1”
  • author’s name cannot be Googled or found elsewhere on the internet
  • author has no photo or uses an AI-generated headshot
  • author bio is extremely vague, vapid and full of soulless buzzwords
  • book previews show lots of AI-generated images, large-print text and wodges of text that seem flavorless and generic
  • no publisher name, or “independently published”

You may use the AI-generated book listings I’ve linked to above as training material. Once you’ve perused those Amazon pages and gotten familiar with the nuances of AI-written works, you’ll probably be able to recognize others in the future.

Also, compare any of those uncanny offerings to a human-written book, from a real-life, talented author. You can verify that an author is legitimate by:

  • Reading the author bio and searching the web to confirm some details
  • Performing a reverse-image search on their bio picture — real authors often reuse their best headshots throughout their digital footprint
  • Finding the author on social media or library websites
  • Shopping for the book on other sites, new or used
  • Finding the publisher of their book, under “See All Details” of the book listing

Final Comments

Remember, you cannot always trust in reviews on Amazon. Certainly, check them out and you may glean some useful information. But many AI-generated book “authors” have gamed the system, and their titles show a lot of manufactured 5-star reviews.

If you think you’ve purchased a book from Amazon that may have been written by AI, you can certainly contact their support agents about it. I don’t see anything in their return policies that addresses this sort of product. But if you chose to seek a refund, using phrases like “failed to disclose AI-generated content” and “deceptive business practices” will go a long way. I doubt they will stop selling an AI-written title, though. Unless you can convince them that it has caused tangible harm.

Many people often urge others to “Shop Local” versus shopping online. In a similar spirit, I hope this post may help some of you to “Shop Human” as opposed to giving money to these shysters and opportunists.

I’ll Pass on Copilot

Copilot is Microsoft’s take on the AI craze. If you use a Windows computer, you’ve probably noticed that Copilot has silently added itself to your taskbar and/or web browser. But not everyone appreciates or cares for this AI tool. If you’re saying, “I’ll pass on Copilot,” here are some steps you can take to get it off your plate.

Quick and Easy Methods

If Copilot is on your Windows Taskbar, you can remove its icon:

  • Right-click your taskbar and left-click Taskbar Settings.
  • Look for Copilot (Preview) and click its toggle switch Off.

If Copilot is barging in on your Microsoft Edge browser:

  • With Edge open, click the 3-dots button to the upper-right, and then click Settings.
  • On the Settings list to the left, click Sidebar.
  • On the right, under App Specific Settings, click Copilot.
  • Click the toggles next to the Copilot options to turn each Off.

You may have to repeat these steps in the future, as Windows Updates may re-enable Copilot, without warning.

More Advanced Tactics

Copilot is treated as a component of Windows, and as such, cannot be removed. But it can be strictly disabled, using stronger methods.

The Group Policy Editor can be used to hamstring Copilot, but that is only available to those running Educational or Professional versions of Windows. Sorry, Home users!

Power users can modify the Windows Registry to disable Copilot. But I discourage this for all but the most tech-savvy. Registry changes are not for the faint of heart and are risky.

If you’re open to using 3rd-party software, I’ve tested DoNotSpy11 and found it to be straightforward and malware-free. This app allows people to disable Copilot and other intrusive features in Windows. If you’re the tinkerer-sort, check this freebie out!

Lastly, you can always switch to a non-Microsoft browser to sidestep Copilot and similar Microsoft distraction. Besides Google Chrome and Mozilla Firefox, there are many solid options out there, including the Brave Browser and the almost-ready-for-primetime DuckDuckGo Browser.

DALL-E

DALL-E is a groundbreaking AI tool that is now open for the general public to use. Made by OpenAI, DALL-E uses artificial general intelligence to create things in stunningly human ways. And in almost no time at all. New users can sign up at the OpenAI website for free.

Creations

Most notably, DALL-E can create a photo or piece of art based on a typed description. Your imagination is the limit, you can ask it for a “bengal cat eating a banana” or “the grand canyon with a rainbow overhead”, and see what AI can create within seconds.

These people do not exist, they were made up by DALL-E

DALL-E also offers a “playground” where it can write something for you. You may type in any kind of writing prompt or topic, and in seconds, the AGI will churn out a very fluent chunk of English text.

In seconds, DALL-E composed a short speech that would take a human half an hour or more.

Caveats

If you experiment with DALL-E, you may notice some creepiness with the images it devises. Keep in mind that this is a work in progress, and also that the AGI is programmed to avoid making human faces that resemble real people. It won’t create anything that looks like you or the President or your favorite celebrity.

OpenAI does restrict this tool from creating objectionable content. Check out their content policy if you have any questions or concerns.

Even if you don’t aim to use this tool, it does have ramifications for our future. While OpenAI is closely supervising and limiting how their tech is used, I do worry about its misuse. I can see how DALL-E and similar tools might contribute to the Era of Deepfakes, if not properly policed. And the writing tool could be abused, but I don’t know what to call it yet. Whatever the term, it’s adjacent to plagiarism.

As you peruse the internet, stay dubious, my friends.

© 2024 BlueScreen Computer

Theme by Anders NorenUp ↑