Stories
Slash Boxes
Comments

SoylentNews is people

posted by hubie on Thursday May 11, @11:54PM   Printer-friendly
from the you-get-an-LLM-and-YOU-get-an-LLM dept.

Pearson has already sent out a cease-and-desist letter over use of its intellectual property:

Textbooks giant Pearson is currently taking legal action over the use of its intellectual property to train AI models, chief executive Andy Bird revealed today as the firm laid out its plans for its own artificial intelligence-powered products.

The firm laid out its plans on how it would use AI a week after its share price tumbled by 15% as American rival Chegg said its own business had been hurt by the rise of ChatGPT.

Those plans would include AI-powered summaries of Pearson educational videos, to be rolled out this month for Pearson+ members, as well as AI-generated multiple choice questions for areas where a student might need more help.

Bird said Pearson had an advantage as its AI products would use Pearson content for training, which he said would make it more reliable.

[...] Bird also said it was usually easy to tell what a large language model such as ChatGPT has been trained on, because "you can ask it".

Bird also sought to point out a difference between Pearson and Chegg, which focuses more on homework assistance.

"They are in a very different business to us," he said. "We see a great differentiator between what Chegg are offering and what Pearson+ are offering.

"We're in the business of helping you learn and improve your skills, not in the business of answering."

He added that - as Pearson was in the business of learning - its products would be hard to replace.

"If all we had to do was read a set of facts in order to learn, there'd be no need for schools, colleges and teachers."


Original Submission

This discussion was created by hubie (1068) for logged-in users only. Log in and try again!
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
(1)
  • (Score: 1, Offtopic) by krishnoid on Friday May 12, @12:08AM

    by krishnoid (1156) on Friday May 12, @12:08AM (#1305972)

    "If all we had to do was read a set of facts in order to learn, there'd be no need for schools, colleges and teachers."

    Ok then -- if you're so smart, *you* figure it out. Legal threats [schlockmercenary.com] are as good place to start as any, let's see how that goes.

    Labs! You also need labs.

  • (Score: 5, Interesting) by Snotnose on Friday May 12, @12:49AM (3 children)

    by Snotnose (1623) on Friday May 12, @12:49AM (#1305980)

    I went to college in the 80s, textbooks were 2/3 my cost. It pissed me off to no end that the used books were 99% the same as the new editions, the only difference being they moved chapters and problem numbers around.

    Undergrad Algebra, Calculus, diffy Q's, Chemistry, Physics, English, and $diety knows how many other disciplines haven't changed in the past 50 years. There is no reason there can't be an open source PDF textbook for these classes, where it costs $5-$10 to print and bind them.

    Fuck textbook publishers, they are a parasite on higher education.

    I remember 1 English teacher who assigned a dozen books, all for $4-$20 at the local mega-bookseller. Problem was, he only used at most 1 chapter of each book. I remember 1 book he used 1 paragraph. That guy was an asshole.

    --
    I just passed a drug test. My dealer has some explaining to do.
    • (Score: 4, Interesting) by istartedi on Friday May 12, @02:02AM

      by istartedi (123) on Friday May 12, @02:02AM (#1305987) Journal

      I went in the late 80s and early 90s. Textbooks were expensive but by no means 2/3 of the cost. The book store would buy them back for 10% and sell them at something like 80% of new when used. It was a minor league grift compared to now. They hadn't invented Internet lock-ins yet. Listen to what the kids are saying, however bad your 80s/90s textbook experience, theirs is almost certainly worse. Of course the weaponization of online courseware pales in comparison to the financial WMDs that student loans are. We generally didn't have those either, and tuition was lower because the colleges knew that most students weren't taking out loans. Once financing becomes the norm in any business, costs always rise and the pay-as-you-go customer gets screwed because it's no longer "What can you afford", it's now "What payment can you afford", and when that payment is delayed, Whoooah, boy. No wonder it's so obscene. As unfair as it seems to let some of those loans go, the more I think about it the more I lean towards saying that it serves the bond holders right if they have to take pennies back on the dollar. These were predatory loans. Thing is, a good chunk of them are held by the government. Yes. We the People. We did it to ourselves. Stop hitting ourselves. Why are you hitting yourself? Hey look, there's something on your shirt...

      --
      Appended to the end of comments you post. Max: 120 chars.
    • (Score: 2) by looorg on Friday May 12, @02:07AM

      by looorg (578) on Friday May 12, @02:07AM (#1305989)

      It's quite annoying when you get those very expensive books and they only use like a chapter or so. Those books are just made to copy, legal-smegal. Still nothing stops you from reading the rest of the book to. To get some actual value out of it. It should give you extra knowledge and context on the matter being described. This whole idea that you should just read certain pages, paragraphs or chapters is borderline idiocy. It's a book. You start at the start and read to the end.

      I recall a professor that wrote a lot of books. So first he got paid by the university as a professor and to teach the class but he also used only/mainly his own books as literature for his classes so he also got paid by the students that had to buy his books. That is some evil academic genius level of thinking and planning.

    • (Score: 2) by aafcac on Friday May 12, @12:56PM

      by aafcac (17646) on Friday May 12, @12:56PM (#1306055)

      In all fairness, the classes were a lot less expensive back then and printing books was proportionally more expensive. The typsetting alone was a lot more work than it is now. When it came to the second hand market, there were fewer options.

      As far as opensource books go, that is a thing, a bunch of the classes that I'm taking currently use ones from openstax. In fact, before my mother retired, she was on a team that used grant money to write one that's now freely available to use for pretty much any purpose and can be remixed if you like. It can even be printed and bound to be like a regular book. It's been a bit since I check, but you could get a hardcover textbook custom printed with color for under $40. Which kind of makes sense. A lot of the cost is royalties and the logistics of storing and transporting large numbers of books. A one off book is in a sense more expensive to print, but with modern printing technology, even that isn't what it used to be.

  • (Score: 3, Insightful) by looorg on Friday May 12, @02:15AM

    by looorg (578) on Friday May 12, @02:15AM (#1305991)

    Even tho I kind of hate big publishers. They sort of have a point. Why should the AI companies and trainers somehow get to train their AI for free on the knowledge gathered and produced by others? Without even referencing where they got the information from. Or I guess they could at least buy one copy of the book. Even tho they would probably be mad then to as it would just be one book, duplicated and replicated by a bot and then the content being distributed to all their paying clients without them getting anything.

    There are a lot of other examples for this I guess in how these search companies, that are now also the AI companies, just basically "stole" the work of others. Indexed it and kept most, if not all, of the revenue while the creators get a pittance at best. Search companies, and now AI companies, are basically information parasites.

  • (Score: 2) by sonamchauhan on Friday May 12, @03:50AM (2 children)

    by sonamchauhan (6546) on Friday May 12, @03:50AM (#1305997)

    Can you actually ask ChatGPT if it's been trained on say, a Pearson Calculus textbook?

    https://www.pearson.com/en-us/search.html/Mathematics/Calculus [pearson.com]

    Bard (https://bard.google.com/) does not give a straight answer:

    QUESTION: As a large language model, does your training data include the book "Calculus" by William L. Briggs ?

    ANSWER: I do not have enough information about that person to help with your request. I am a large language model, and I ...

    • (Score: 0) by Anonymous Coward on Friday May 12, @03:50PM (1 child)

      by Anonymous Coward on Friday May 12, @03:50PM (#1306115)

      Are you expecting that ChatGPT would say yes and provide enough word for word exact samples to prove it was trained on the stuff?

      What if it doesn't say yes but it was actually trained on it?

      It's often confidently wrong about stuff too...

      • (Score: 3, Interesting) by sonamchauhan on Monday May 15, @10:15AM

        by sonamchauhan (6546) on Monday May 15, @10:15AM (#1306356)

        Those are my questions too.

        The Pearson CEO is confident he can interrogate ChatGPT and have it confess to all his products that it ingested. ("You can ask it", he says.)

        My question is how? At what point do ChatGPT's pontifications on generic Calculus theory map to material lifted from his specific text.

(1)