Chat GPT and Data Privacy

Big companies are just trying to sell you stuff.
December 26, 2024 by
Chat GPT and Data Privacy
Andrew Valenzuela
| No comments yet


TLDR

  • AI privacy: ChatGPT uses data for model training but doesn't compromise your personal privacy directly.
  • For complete control: Use offline LLMs for sensitive tasks.
  • Big companies like Google, Microsoft, and OpenAI use your data to sell products, not spy on you.
  • Advertising IDs track your behavior anonymously to target you with ads across devices and apps.
  • Operating systems (iOS, Android, Windows) collect data, even with GPS and location off.
  • Limit data sharing: Use ad blockers, Firefox, separate profiles, or switch to Linux.
  • For maximum privacy: Consider privacy-focused OSes (Tails, GrapheneOS), Tor Browser, and VPNs like Mullvad.


Don't Be Evil; Sell More Shit!


Large companies don't care about your data, in the regard that they don't care to manipulate you.  Google isn't listening in on your private conversations and Microsoft isn't taking snapshots of your computer screen.  

These big companies are collecting your data.  But they aren't collecting and using it in the way that you or I would.  There's no personal vendetta or blackmail going on here.  

They are up to something far more nefarious!  

Selling you shit!  Lots of if!

OpenAI (the company behind ChatGPT), and the like, are using your data in the same way. 

That is capitalism baby!  

"Build a better mouse trap."  

(well, It's more like:  Build a half decent mousetrap with planned obsolescence in mind, so you can sell more mouse traps.)


Let's give ALL of our data away!!!

​You already are, here's how.

We've all had that weird scenario.  

Talking to some stranger at the store about a new widget (product) you've never heard of.  
Then BAM!  Facebook hits you with an ad for that widget you hadn't heard of before.

Obviously your phone is listening to your intimate conversations. (sarcasm, that's me being sarcastic)

What the heck is an advertising ID?

  • iOS
  • Windows
  • Android
  • MacOS
  • (every non-FOSS operating system)

Are invasive operating systems that sell and trade your data with companies in order to sell you stuff.  All of which is done ANONOMOUSLY via an "Advertising ID".  

Advertising IDs are anonymous and re-settable unique identifiers used to target you and track your interactions.  Important to know, these don't have any personally identifiable information in them.  But they are unique to your device / account.  Companies (like Microsoft) get this advertising ID and use it to track your interactions and interests.  They build a profile based on this information and then use it to sell you stuff.   

Andoid Phone Ad ID:


Let's go back to the convo you had about that new widget.  Obviously your phone had to be listening into your conversation.  To show you an ad about that new widget you never heard of....

Which would mean:

  • Real-Time voice to text (that is always on and always working and transmitting data for Billions of users simultaneously)
  • Real-Time translation for all of these convos (we speak Spanglish in my area.  And in Houston there's something like 20 different spoken languages)
  • Constant data transmission of all this data (it's not much individually.  But for Billions of interactions this would become a huge amount of data).  

Or, the easy way:

  • Each app requests access to your Advertising ID, and other available data (like your general location, which is determined by your IP address, MAC address, or other info).
  • Use this data to map an interaction:
    • User 1​ is searching for widget​ and interacting with it frequently.
    • User 2​ was in the general area that User 1​ was.
      • Send User 2​ an add for the widget​.



Oh NO! I'm being tracked, help!

There are too many ways to track you, and your activity, for me to list here.  

What you should know, we can get your:

  • General Location (even if you have GPS off)
  • Device Info:
    • Make/Model 
    • Sensor data
  • Carrier/ISP data
  • Your behavior within an app/website (what you click on, how you've hovered)
  • Facebook (Meta) is always tracking you, even if you don't have an account (Meta Pixel)

Your advertising ID isn't limited to being on your device btw.  It's tied to your account (oulook.com, gmail.com, apple.com, facebook.com).  All of these different sites (if it is free then you are the product) are using your data to sell you more stuff.  It's also just one of many things which is tracking your info.

What this means is that anytime you visit a site which uses Google (Alphabet), Amazon, Facebook...  Your data is logged and tracked.  

These large companies use this data to send you more relevant ads...

They are not the ones you need to worry about.  It's the individual companies that are requesting your information which should concern you. 

Story Time

I recently heard a disturbing ad, which contained a warning, masquerading as a joke.  It was in a political podcast. 

"Sign up for our political podcast special membership.... And remember, we know if/how you vote!

That wasn't a joke.  Voter roll calls are publicly available information.  They cannot tell who you voted for (as that is anonymous).  But they can tell if you voted.  
And by logging your various identifiable information.  Cross referencing your user name across social media sites.  
They can build a reasonably accurate Personal Profile​ about you.  

Should I be concerned about Google and Amazon having my data???

No, because their motive is to sell you more stuff.  

Should I be concerned about PoliticalPodcastNetwork.evil having my data???

Yes, because their motive is to control how you think/vote.  

Bond villains don't exist

The idea that a large company like Google or Facebook would collect data to control the world (or enact a personal vendetta) is just ludicrous.  

That's bond level type evil that would require an unrealistic level of competence and cooperation.  

Such as:

  • Having full control over a popular social media platform that is already tied into every service.
  • Owning & Controlling a telecommunications provider service (ISP).
  • Getting the cooperation of, not one, but multiple political entities to control laws and skirt current laws.
  • And to do all of this you would need immense wealth and popularity, not to mention a huge and powerful disinformation campaign to drown out the Canaries in the Coal Mine.

That is just silly.  But, small entities may still use your data against you.  So be mindful what you share.


How to limit the data you are sharing.


How to be as anonymous as possible.

To do this, YOU have to change.  There is no way to protect you from your own behavior.

If you have a real need to remain anonymous online.  You should contact a professional like me.  Being private and safe isn't simple.  

Here are some jumping off points:

  • Use an operating system that is dedicated to privacy:
  • Use an anonymous browser:
  • Switch to a de-Googled mobile OS:
  • Use a VPN (one that doesn't advertise, if it advertises, its a money hungry company, and is probably a company willing to comply with authorities and stores customer data) (also, VPNs don't do what the companies advertise that they do.):
    • IVPN
    • Mullvad
    • VPNs don't do what "nordvpn" claims they do.  It's good to claim a different country, and to protect you from public wifi vulnerabilities.  Outside of that it is not the "privacy shield" it's advertised to be.  Texas Technician uses its own VPN to protect against public Wifi vulnerabilities (which if free to setup and included on any high-quality router).



ChatGPT, AI and Privacy

Ethics and customer protection is builtin.  Safeguards are put in place for a reason (because people are not safe):

Even so, you shouldn't share anything with ChatGPT, or any AI/LLM that you wouldn't give to your secretary.  

Keep in mind that your data is likely being used to train and improve the model.  But it's practically nothing.  

Meaning That:  
YOU are NOTHING to this giant corporation, whose main goal is to make as much money as possible.  

If you're not using AI/LLMs, you will fall behind.  This would be like sticking to paper/pen/and a ledger book, because you don't trust Spreadsheet programs (Excel). 


Switch to a Local (Offline) LLM

There are so many out there.  I use a few for in-house purposes. 

Here is a video about an easy to install local AI I have found:  

Alpaca, use opensource AI (offline).


Chat GPT and Data Privacy
Andrew Valenzuela December 26, 2024
Share this post
Tags
Archive
Sign in to leave a comment