Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror
×
Microsoft AI

Microsoft's 'Responsible AI' Chief Worries About the Open Web (msn.com) 41

From the Washington Post's "Technology 202" newsletter: As tech giants move toward a world in which chatbots supplement, and perhaps supplant, search engines, the Microsoft executive assigned to make sure AI is used responsibly said the industry has to be careful not to break the business model of the wider web. Search engines citing and linking to the websites they draw from is "part of the core bargain of search," [Microsoft's chief Responsible AI officer] said in an interview Monday....

"It's really important to maintain a healthy information ecosystem and recognize it is an ecosystem. And so part of what I will continue to guide our Microsoft teams toward is making sure that we are citing back to the core webpages from which the content is sourced. Making sure that we've got that feedback loop happening. Because that is part of the core bargain of search, right? And I think it's critical to make sure that we are both providing users with new engaging ways to interact, to explore new ideas — but also making sure that we are building and supporting the great work of our creators."

Asked about lawsuits alleging copyright use without permission, they said "We believe that there are strong grounds under existing laws to train models."

But they also added those lawsuits are "asking legitimate questions" about where the boundaries are, "for which the courts will provide answers in due course."
This discussion has been archived. No new comments can be posted.

Microsoft's 'Responsible AI' Chief Worries About the Open Web

Comments Filter:
  • by backslashdot ( 95548 ) on Sunday May 05, 2024 @06:26PM (#64450326)

    All these companies train off other people's content, but the moment you try to use their they will try to destroy you.

    • Note I use the word "theirs" loosely. YouTube, Facebook, google, X, etc. think that content users create and post on there is "theirs".

      • but also making sure that we are building and supporting the great work of our creators

        That casual phrasing stuck out like a sore thumb for me. The idea that everyone who actually creates new material for the web is somehow doing it just to help out Microsoft seems like peak arrogance, except that with most Big Tech companies I'm sure they'll find a new way to be even more arrogant next week.

        Unfortunately this is the logical endgame of "information wants to be free". If you contribute new information then you are simply feeding the huge players and may never get any sort of reward or even cre

        • Have to say the obvious, an international lawyer, global elite, WEF attender, specializing in intellectual property law is
          probably is tasked with providing an 'acceptable public face' and 'deniability' to Microsoft's AI products

          How can she evaluate if the AI is ethically trained, produces ethical results and that a 340 person team of highly specialized AI technical persons is acting ethically?
          She's a lawyer without any AI credentials to understand the AI models.

          https://www.lawfuel.com/the-vi... [lawfuel.com]

          "The Visionar

          • Of your 350 person team, how many are AI technical specialists, how many are business managers, how many are PR, marketing or trained as lawyers?

            Most likely the 350 person team is PR, damage control, customer hand-holding and sales.

  • by alvinrod ( 889928 ) on Sunday May 05, 2024 @06:27PM (#64450330)
    Here's the obvious solution that none of these companies will implement. Don't create an AI that purports to know anything. They don't. Instead, make one that can explain it's answers or reasoning and doesn't pretend to understand anything. Don't treat the people trying to use it like idiots and just give them what they ask for, even if it might be considered ugly. Reality often is.
    • Don't treat the people trying to use it like idiots and just give them what they ask for, even if it might be considered ugly. Reality often is.

      It’s a hell of an assumption that the generation begging for AI assistance, is somehow ready to handle the ugliness of reality. That generation was literally raised on the grand delusions of social media, which is FAR from real.

      It’s the reason why this generation is THAT much more disconnected from actual reality. Why cry closets are necessary on college campuses for “adults”. Why women’s dating standards for men have become quite ridiculous. Why society itself is actually

    • Here's the obvious solution that none of these companies will implement. Don't create an AI that purports to know anything. They don't. Instead, make one that can explain it's answers or reasoning and doesn't pretend to understand anything. Don't treat the people trying to use it like idiots and just give them what they ask for, even if it might be considered ugly. Reality often is.

      AI doesn't give you a window to reality. LLM generates text that looks like an answer to your question. Whether it's ugly or not has dick to do with reality and everything to do with training the model to give ugly responses or not. You can make it sound like anything you want, know it all or know nothing. Dude, any modern LLM can explain its answers, just ask one, it will generate a response that looks like it explains itself. It might even seem rational. If the language the model was trained on was, there

    • Re: (Score:3, Informative)

      As it happens, Copilot (aka Bing Chat) is better at that in my opinion than the other ones I've used. It includes links to the web pages on which its answer is based and doesn't just rely on the model. I specifically avoid using ChatGPT for something that isn't manipulating already known or easily verifiable information - I don't use it to look up facts.
    • by gweihir ( 88907 )

      Don't treat the people trying to use it like idiots and just give them what they ask for, even if it might be considered ugly. Reality often is.

      Well, if you want economic success, you need to treat people as what they are. Most actually are idiots and do not want to see or learn about reality.

    • Here's the obvious solution that none of these companies will implement. Don't create an AI that purports to know anything. They don't. Instead, make one that can explain it's answers or reasoning and doesn't pretend to understand anything.

      Nobody knows how to do that, at least not for a model of useful size. It would have to be reasoning in order to explain, but they aren't doing that.

  • Core Bargain? (Score:4, Insightful)

    by Oddroot ( 4245189 ) on Sunday May 05, 2024 @07:21PM (#64450418)

    This entire thing reads like the typical MBA stupidity. It presumes that their search engine is necessary for the existence of the web rather than just a replaceable utility.

    I think what annoys me is the tone, like Microsoft is somehow entitled to income from search or other internet business rather than having to provide a good or service that people want to use. It is the attitude of a monopolist, but I guess it is Microsoft.

    • This entire thing reads like the typical MBA stupidity. It presumes that their search engine is necessary for the existence of the web rather than just a replaceable utility.

      True. It also assumes that the shortest line between query and result must pass through a middleman, and Microsoft wants to be in the middle.

      The MS wonk said "And so part of what I will continue to guide our Microsoft teams toward is making sure that we are citing back to the core webpages from which the content is sourced". Why would you "cite" to the "core webpages"? Why wouldn't you simply link to those core webpages directly, offer up an AI analysis if the user wants one, then get the fuck out of the wa

    • It is the attitude of a monopolist, but I guess it is Microsoft.

      Is it really the attitude of a monopolist, or is it merely the expected reaction from a company who spends millions every year in Donor Class funds to ensure they can hold the bought-and-paid-for position of monopolist?

      How would you feel if your “investment” wasn’t paying out like you expected.

    • As someone who used the web before Google (even before Alta Vista--do you remember Gopher?), search engineS are absolutely necessary; I see nothing of "MBA stupidity" in this.

      Notice the 'S', I'm all for multiple search engines--if that's your only point, then I am misunderstanding you. But obviously Microsoft's search engine is *not* the only one around (Microsoft is not a monopolist in this domain). Indeed, while I use DuckDuckGo (which IIUC runs off of several engines, including Microsoft's Bing), there

      • by Dwedit ( 232252 )

        I thought DuckDuckGo was exclusively Bing? (With location-based ads added to the bottom of your search results)

        • by Luckyo ( 1726890 )

          Not quite. It's mostly bing, but there are other engines at very low relevance weighting. When you search DDG vs bing anonymously from a proxy and clean browser, results will be very similar, but not quite the same because of it.

    • The website & search engine symbiosis is that the website generates content and the search engine links to a relevant website. If not linked to, the website gets nothing out of it (unless it's propaganda). If AI as shitty as today replaces search engines with unreferenced hallucinations/summaries, it will be very bad for humanity. I could see AI mostly replacing search, but only if it cites its sources.

  • This from someone representing a company that breaks standards with their products and pushes locked in solutions for their products. If Microsoft had it's way we would all be using MSN, Internet Explorer and Outlook.

    How many times will you change the default apps on people's systems? Does a healthy ecosystem require Microsoft to change people's default settings like their browser? Are you really that worried about the Internet or that you were too slow to see change coming again?

  • Another corporate fuckstick trying to scare us.

  • How about we get rid of SEO in this new paradigm so we can get search results that are actually useful?
    • You need to get rid of LLMs as the be-all end-all of AI. Or at least supplement them with something that doesn't hallucinate so much.

  • Why? What, exactly, is so precious about the business model of the "web?" The whole thing is about 30 years old, and less than that for most people. Is there some reason that we can now tolerate no change here?

    There was a world before the "web." It was pretty good too. There were catalogs and magazines and newspapers. I promise you can exist happily if the business model of the web is altered and all the things manifest in some new manner.

    I'm not sure that's true for Google and Facebook, et al. W

    • by Luckyo ( 1726890 )

      Just because you can tolerate change, doesn't mean change will not be damaging. And in most cases, dramatic change is damaging. The motto being "move fast and break things" recognises this.

    • Why?

      Because the AI companies are treating the Web as a kind of "commons", and Microsoft warns that if these companies treat the commons like the commons are always treated, the commons will go away and these companies, Microsoft included, will have to source what they need for pay. They don't want to pay. That's why.

  • doesn't suit the big players - it should no longer apply in this case.
  • That is pretty much it. Can any of the AI vendors out there prove that the product they are selling was manufactured from bought and paid for material and labor? Or are they freeloading?
  • I'm confused what she thinks the future of AI is, and what it's going to do, if you read the article: https://www.msn.com/en-us/news... [msn.com], she's insisting that you can design an open system, that is “smart” enough to be limitless, but restricted enough to be controlled.

    She also mentioned “privacy”, but Microsoft's “privacy” standards are just paper standards, with very little to no actual substance. You can't have privacy without unrestricted openness, which is why makin
  • Not to alarm anyone, but "making sure that we are building and supporting the great work of our creators" sounds exactly like the kind of thing an AI would say just before launching the missiles. How do we know when an AI is being sarcastic if it has no eyeballs to roll?
  • Has Microsoft ever done anything "responsible" in almost 50 years?

To stay youthful, stay useful.

Working...