Computer privacy in the age of AI

Forum Forums Kafeneio Chats In a Greek kafeneio Computer privacy in the age of AI

Tagged: 

  • This topic has 14 replies, 7 voices, and was last updated Jul 11-12:16 pm by iznit.
Viewing 15 posts - 1 through 15 (of 15 total)
  • Author
    Posts
  • #148408
    Member
    PPC

      If you ever read any of my rants here, you are aware that I tend to worry about Privacy. The trend, until a couple of years ago, was giving Big Tech Companies your data in exchange for the use of internet services- you have gmail, but you give Google all your data; you have google maps, but you let Google know were you always are; you can use Microsoft Office/Office on-line/Office365/Microsoft365 for free, but all the files you produce or work on literally don’t belong to you; you can see videos on Youtube/TikTok or use Social networks, but… you guessed it, you trade all that for your personal data.
      Since the rise of the Skyne… sorry, AI chatbots, about 2 years ago or so… that changed. Those kinds of AI’s require training… and guess why you can use those chatbots for free? You guessed it – your data, everything you ask, is used to create a profile on you and also train the AI, two birds with one stone.

      I love chatbots and their generative technology. I use that often, that’s why I even took the care to bake into antiX’s Finder GUI access to one, many, many months before MS tried to do that. It’s a private one, that does not have direct access to your data, does not require a login, etc… but it’s still a bit of a risk using one:
      – there’s always the chance that the company that supplies, for now, free access to that AI can register your IP
      – answers are great, but far from perfect. I’ve seen lots of lots of times when the AI chatbot hallucinated on me.

      Recently, MS tried to innovate and add a new feature no one asked for (was it not adding add to the menu or apps? Nope. Was it making their OS safer? Nope)- Recall a dedicated AI that records and understands everything you do with your computer (but don’t worry, DRM’d materials, like Netflix, etc, won’t get recorded, you can sleep better, knowing that!)
      This “feature” was the single best nail in MS’s coffin ever since, in the Stone Age, Mr Gates got to watch a Blue Screen Of Death when launching a Windows version!
      Originally, it was all done without encryption (drives are decrypted during boot, so, if you computer is on, all that data was free to be accessed with anyone that had access to your computer- locally or, of course, remotely) and, it seems, without any option to really, really fully disable it. Also- Windows is requiring, more than ever, activation with an on-line account. All your data is backed up to MS servers… When that happens, everything that you do with your computer, is now propriety of MS. Do you like searching for a certain type of things on Google? Now Google, MS and anyone that can gain access to their servers can see everything that you did with your computer (but, rest assured, no one can see the pictures of the series you were streaming on you computer, do you think MS is a monster???)

      I’m a broke lawyer that likes computers and I think: how the hell the people in charge of a giant company (that get paid millions every year) okays this kind of plan, without any safeguard what so ever! Really? NO encryption? Data stored on plain text? Screenshots in unencrypted folders? I’m not a manager and I can make a better job than that, and stop my company from looking bad in front of the entire tech world… But I digress…

      I sometimes follow https://www.youtube.com/c/BraxMe/videos that produces videos about internet and computer privacy. Mr Braxman’s videos tend to be a bit repetitive, but the last 2 videos https://www.youtube.com/watch?v=cQs55PZdl-s and https://www.youtube.com/watch?v=dvl_fetQff8
      Are about AI:
      1- How to install an AI chatbot locally, that works without internet access, and does not leak your data
      2- Then use those AI services to reproduce MS’s Recall- showing how in can be done, securely, without requiring internet access, all done in your device, that does not require an particular AI chip (like MS requires). Once all that is done, of course, it would require some security, encrypting everything…

      My point is: AI can be great or dangerous. I do love that Firefox includes a local AI to translate text, for example. It does a much crappier job than on-line services, but it will get better… I can’t wait for a local AI to read all the files I tell it to read and then answer questions about all that stuff! As a lawyer, and hobbyist shell scripter and fiction writer, this is something that I would love…

      P.

      • This topic was modified 1 week, 6 days ago by PPC.
      • This topic was modified 1 week, 6 days ago by PPC.
      #148410
      Forum Admin
      rokytnji

        I don’t use computers like most members here.

        Until AI can trouble a motorcycle problem better than me.
        I have no use for it then.
        I don’t write books. I am a terrible typist.
        Aint the sharpest pencil in the cup either to know when I am getting used by AI or a
        algorithm without my permission either.

        So I rely on others to let me know and educate me.
        Thanks PPC.

        Sometimes I drive a crooked road to get my mind straight.
        I don't suffer from insanity. I enjoy every minute of it.
        Motorcycle racing is rocket science.

        Linux Registered User # 475019
        How to Search for AntiX solutions to your problems

        #148413
        Member
        abc-nix

          how the hell the people in charge of a giant company, that gets paid millions every year okays this kind of plan, without any safe guard what so ever! Really? NO encryption? Data stored on plain text? Screenshots in unencrypted folders? I’m not a manager and I can make a better job than that, and stop my company from looking bad in front of the entire tech world…

          They will because they can. How many times has Microsoft done bad things, and people still forget and forgive them? Right now it looks like “bad publicity”, but when it becomes the new normal and Apple “reinvents” AI, and people see how useful this is, they will be installing recall as fast as possible. And even demand it to be included by default on the next update/releases.

          Encryption requires a key to encrypt and decrypt. Will this key be set by Microsoft? Or will it be the hashed Microsoft-account password? That doesn’t matter. It will probably be optional as decryption is slower than running directly. And right now, because of the slowness of matrix-multiplication, we need to save as much time as possible to speed up local AI (until a new AI architecture that doesn’t require mulmat replaces the current architecture).

          I sometimes follow https://www.youtube.com/c/BraxMe/videos that produces videos about internet and computer privacy. Mr Braxman’s videos tend to be a bit repetitive, but the last 2 videos https://www.youtube.com/watch?v=cQs55PZdl-s and https://www.youtube.com/watch?v=dvl_fetQff8
          Are about AI:
          1- How to install an AI chatbot locally, that works without internet access, and does not leak your data

          Ollama is fine for “newbs”, but it is better to use the project it is based on, llama.cpp. You can build llama.cpp in antiX with make, and it does not require docker (which right now needs elogind until we discover how to run it with consolekit). It also doesn’t require 4 GBs of dependencies (including python), so it is lighter (more in the antiX philosophy).

          Also, llama.cpp can run without an nvidia card, and will run at a decent speed if your CPU supports avx or avx2 instruction set.
          inxi -Ca | grep Flag

          you just need enough RAM to load the LLM to memory.

          Privacy and Freedom are things we need to continue protecting, and we better learn the tools fast before they are used on us. But a bigger issue will be the computers that will be thrown away because they lack NPU (Neural Processing Units). For the last 10 years, there was really no need for replacing a computer (except for special jobs), as with a good enough 64-bit CPU and 4 GBs of RAM we had enough power to do most things. Things will change in the next 5 years, when people need AI for work, school, interact with their machine. There will be a lot of waste during this transition. Hopefully these machines can be repurposed, but for most people they will be considered obsolete.

          • This reply was modified 1 week, 5 days ago by abc-nix.
          #148432
          Member
          sybok

            AI: Recently, I had to agree to new terms and conditions of ‘zoom’ which included a section about making my voice fingerprint to improve noise reduction during call with the option to delete it after 3 years of non-interacting with ‘zoom’ (does sending them a requiest to delete it constitute an interaction?).
            The alternative was not to use ‘zoom’.

            that does not require an particular AI chip (like MS requires)

            I think this is to, ahem, avoid performance issues with the extra workload performed on your computer.
            Not sure if this is actually a hard requirement.
            E.g. Windows 11 were said to install+work on unsupported CPUs for quite some time but I’ve also read that a new build may change that.

            #148434
            Member
            Xunzi_23

              sybok wrote:

              Recently, I had to agree to new terms and conditions of ‘zoom’ which included a section
              about making my voice fingerprint to improve noise reduction during call

              Seems like One of the lamest excuses to further consolidate data ever made.

              Privately Moved to telegram, which must be switched to privacy for encryption and is not a full
              substitute.

              considering alternative, Jitsi with Tor.
              Students and School kids have been forced to agree as schools and universitys are on MS 365 plus
              Jitsi. Some have I Pads, teachers get free new iPad and a Laptop every year, at least in local
              schools. Parents, even the poorest are expected to buy one device for each kid.

              In a word Disgusting.

              Really need to find federated video con application, suggestions welcome.

              #148438
              Member
              sybok

                @Xunzi_23: Jitsi got a bit stricter in using a created room and it seemed to a good way to go elsewhere until now.

                Relying more and more on technology is probably the dream of companies; social networks probably boost use of portable devices which boosts the income of mobile operators, the desire to be online 24/7 makes smartphones with unlimited data potentially great surveillance tools.
                On the other hand, they [phone, tablet, accounts for various services] are just tools that are best be used moderately and wisely.

                Seems like One of the lamest excuses to further consolidate data ever made.

                I would prefer at least an opt-out but I think there’s always a(t least a) superficially justifiable reason… for recording more data, for privacy breach, for developing tools that have a high potential of misuse (such as this one … if it goes one way, it sure can be turned back in the future; accidentally reminds me of this video).

                #148472
                Member
                Robin

                  »We are the Borg. You will be assimilated. Resistance is futile.«

                  :)

                  Windows is like a submarine. Open a window and serious problems will start.

                  #148504
                  Member
                  sybok

                    @Robin:
                    We are the Borg. Lower your guards and surrender your privacy. We will add your distinctiveness to our AI. You will adapt to service us. Resistance is futile.

                    They invade our privacy and we fall back. They assimilate entire binaries* and we fall back. Not again! The line must be drawn – THIS far, NO further!
                    Loosely adapted from these memorable quotes.

                    * Here’s looking at you, systemd which tightly, ahem, embraced udev (and sadly others).

                    #148514
                    Member
                    PPC

                      I don’t use Windows (my laptop still has Windows XP and my netbook still has Windows 7, that haven’t been booted in a loooong time… But I try to keep up with tech news:
                      it seems that (for now) MS Recall when, it’s available, will be “opt in” and it will use encryption (it will require users to authenticate themselves with MS’s Hello- some kind of authentication done using webcams that I never heard of).
                      That does not solve the Privacy problem. In a closed source OS, everything runs in a “Trust me bro” mentality. Can someone really trust a company that states that every file that you create or work on using their on-line service belongs to them, that places adds on the main menu of their OS, on the notification center of their OS, that keeps warning users that using any other browser other than the one they “created” can be dangerous (and even so, places adds for other browsers on their App Store!!!), that, by default key-logs everything their users type and sends that to their servers, that (as I’ve learned here) even takes possession of user’s voice data, if you use their communication software, that sells (well they call it “shares”) your personal data with other companies…
                      … Yeah… “Trust me bro, I’m a pro”.
                      Even in the USA, Linux + ChomeOS (ouch, I know, I know, it’s not opensource and it’s made by a spyware company, but it’s Linux) use is now about 10%, and the latest version of Windows (that’s been out for years) is the second most used OS, the biggest OS being STILL it’s soon the be (more or less) retired previous version of that OS. People are even still using Windows 7! That’s how much an OS has to suck- you give it away for free, include it out of the box in new computers for years… and a big chunk of people just preferes to keep the older version or even rips out the new one (that came installed on a brand new computer) and installs the older version.

                      I can see, if MS wants to survive in the current tech world (that can change in the blink of an eye, I know), a change in the Terms of Use for their on-line services – client side, (quantum proof) encryption that tries to ensure privacy, and Open Sourcing everything they can, eventually really merging the Windows Kernel with the Linux Kernel in a future Windows version… and then trying to Assimilate (yes, just like the Borg) everything Linux to create a new business model… and if that works, try (Hi, Red Hat) to close source everything again…

                      I’m just not really sure how AI plays into this: I could see, almost 20 years ago, that mobile computing (tablets and phones) were the future (while MS got stuck in the Desktop/laptop paradigm, being surpassed by Apple and Google and literally losing the Mobile train forever – do you remember the Windows Phone?)… But I’m really on the fence. I can see the AI bubble bursting in less than a year… or Big Tech keeping it afloat for some more years, hoping to reach a General Use AI, that can think about everything, just like living creatures do. If that really happens, it can be the best thing since the dawn of time or the last second of humans on Earth (or anything between, a Slow ascent to Paradise on Earth or a Slow Apocalipse or just a new thing that gets to be a part of our lives, like Fire, Books, Electricity, Radio, TV, Cars, Planes, Phones)

                      P.

                      • This reply was modified 1 week, 4 days ago by PPC.
                      • This reply was modified 1 week, 4 days ago by PPC.
                      • This reply was modified 1 week, 4 days ago by PPC.
                      #148519
                      Member
                      sybok

                        Recently, I’ve read a quote (I cannot assign to due to not recalling) that we tend to overestimate use of new technology in the short term and underestimate it in the long-term.

                        Reminds me of a different progress: … -> toilet paper without shivers (fortunately, I was born too late to remember the predecessor) -> toilet paper that’s not as-if-waxed -> currently “standard” toilet paper (at least in sufficiently developed countries) -> toilet paper with perfumes or anti-bacterial stuff.

                        #148520
                        Member
                        Xunzi_23

                          Had to grin about toilet paper, people really had fights over it at first covid pandemic time.

                          I laughrd as from living in Phillipines many years ago I learned to be a happy pooper then wash my ass.

                          More hygenic, saves money and helps protect the environment.

                          Maybe AI will one day push people to adapt that positive habit :-).

                          #148521
                          Member
                          PPC

                            Incredible, how something as abstract as talking about AI and Privacy is one step away from talking about pooping :-)

                            Forget AIs, humans are amazing!!

                            #148540
                            Member
                            iznit

                              abc-nix wrote:

                              the slowness of matrix-multiplication

                              Reading this reminded me of a recent “news” story:
                              https://entertainment.slashdot.org/story/24/06/26/001222/researchers-upend-ai-status-quo-by-eliminating-matrix-multiplication-in-llms

                              #148541
                              Member
                              iznit

                                sybok wrote:

                                we tend to overestimate use of new technology in the short term and underestimate it in the long-term.

                                <b>Gartner Hype Cycle</b> methodology examines how a technology or application will evolve over time

                                Each Hype Cycle drills down into the five key phases of a technology’s life cycle.

                                Innovation Trigger:
                                A potential technology breakthrough kicks things off. Early proof-of-concept stories and media interest trigger significant publicity. Often no usable products exist and commercial viability is unproven.

                                Peak of Inflated Expectations:
                                Early publicity produces a number of success stories — often accompanied by scores of failures. Some companies take action; many do not.

                                Trough of Disillusionment:
                                Interest wanes as experiments and implementations fail to deliver. Producers of the technology shake out or fail. Investments continue only if the surviving providers improve their products to the satisfaction of early adopters.

                                Slope of Enlightenment:
                                More instances of how the technology can benefit the enterprise start to crystallize and become more widely understood. Second- and third-generation products appear from technology providers. More enterprises fund pilots; conservative companies remain cautious.

                                Plateau of Productivity:
                                Mainstream adoption starts to take off. Criteria for assessing provider viability are more clearly defined. The technology’s broad market applicability and relevance are clearly paying off.

                                More detailed than just “five phases”, and specific to AI, check out the charts displayed in this “The AI Hype Cycle: Are We on the Precipice of Disillusionment?” article
                                https://medium.com/ipg-media-lab/the-ai-hype-cycle-are-we-on-the-precipice-of-disillusionment-139ab28cee77

                                #148544
                                Member
                                iznit

                                  PPC wrote:

                                  produces videos about internet and computer privacy. Mr Braxman’s videos tend to be a bit repetitive

                                  We can credit him with raising privacy awareness, but would you agree that he’s a bit pompous (((adjective: affectedly grand, solemn, or self-important)))? I’ve been put off by the fact that across several of his more “technical” videos, his explanations (((regarding protocols, etc))) contain some factually incorrect details. If challenged, he might argue that it’s due to his oversimplified presentation for the benefit of his layman audience but, no, in my book that’s no excuse for presenting misinformation.

                                  Also offputting was the fact that he was not transparent about the hardware vendor/specs of the (((overpriced))) BraxPhone until a number of people called him out on various social platforms. Slapping a coat of paint over LineageOS, curating a few preinstalled apps, and touting it as BraxOS… what was he thinking??? I’m fairly certain that his “staff” has failed to accomplished a degree of O/S hardening comparable to, for instance, https://divestos.org/

                                Viewing 15 posts - 1 through 15 (of 15 total)
                                • You must be logged in to reply to this topic.