Global Fascist Technocratic Surveillance State

[cherokeetroy]

I don't see how there's any possibility of escape from the digital prison

It would require major sacrifices that I don't see many people willing to make

We've become fully dependent on technology and it's going to continue advancing

Every material asset is going to be tokenized 

Everything that we own or utilize in the real world is going to have an identifier on the blockchain ledger

In order to access our assets or to buy, sell or perform any transaction relating to our assets, records, money, property, etc...will require a digital ID

Digital ID will be acquired through performing a biometric scan---facial recognition, fingerprint or iris scan and at some point DNA

That digital ID will be your access to participation in their system 

You'll be required to scan your ID to gain access to your records, certificates, documents, banks, government institutions, voting, transportation (public and private), medical services, retail industry, entertainment

Stores, gas stations, facilities, venues, vehicles, etc...will require a scan of your ID, hand, face, eye before entry/access

Every transaction you make in real life/real time will be permanently documented on blockchain 

You will leave a digital trail/footprint wherever you go, whatever you do

Constant 24/7 recorded surveillance 

A global digital prison 

Not in the future 

Now

This should be the most important topic for discussion but the majority of humanity is sleepwalking right into the cage

I made a thread about the new global digital monetary system that's being voted on today in the US

That thread is relegated to the "Off Topic" section

That should tell you where we're at

We're Done

[ArMaP]

These are great. Unfortunately, I don't know how to fight tech with tech as I've always resisted it and know so little about it.

Sometimes, you don't need to fight tech with tech, but you need some knowledge about the tech you are fighting.

Some years ago I created 3 Facebook accounts to help me play games. I gave real information to them, but I didn't give the same information.
For example, for one I listed my favourite movies, choosing only one type of movie.
For another I listed my favourite songs, but only those that were of a type completely opposed to the movies.
I also used variations of my name in other languages and clicked on different types of adverts and chose to ignore other types.
In the end, Facebook reached the conclusion that they were really three different people, even thinking one of them was a gay man from Eastern Europe.

Partial (but true) information may be a strong weapon, but you need to know what the specific tech you are fighting needs, to be able to mess with it.

[ArMaP]

I was reading a Professors' Reddit page and they say that using white text, very small with a white background, where only AI can read/detect it, is one way to identify whether or not you are engaging with AI. Unfortunately, some of their students use dark background mode and were able to read the extra instructions to the AI.

Last month or so I read about a phishing method made to mess up anti-phishing AI agents. It used a specific, specially made, font, in which an A, for example, would look like an M. As AI reads the page's HTML, it read what was really there, with As being As, while humans would read a different thing.

[quintessentone]

Last month or so I read about a phishing method made to mess up anti-phishing AI agents. It used a specific, specially made, font, in which an A, for example, would look like an M. As AI reads the page's HTML, it read what was really there, with As being As, while humans would read a different thing.

I wonder if AI knows pig latin? lol Or the jibberish words I have to type to get into my accounts may be another way around AI where we would all need to make up our own new language and spelling and change it up often. I don't know, just spitballing.

Edit: Yes, AI knows pig latin.

How to foil them with language...

"Natural and Low-Resource Languages
AI models do not comprehend meaning, context, or intent; they process patterns. Consequently, they struggle most with:

• Natural Language (Human Speech): AI does not "understand" language but predicts text based on statistical correlations. It lacks true semantic comprehension, analogy, or metaphor interpretation. 
  
• Low-Resource Languages: Languages with limited digital text, literature, or media in the training corpus (e.g., many indigenous languages with few speakers) are poorly supported because there is insufficient data for the model to learn patterns. 
  
• Complex Linguistic Features: Languages with extensive tonal variations, ambiguous contextual nuances, or code-switching (mixing multiple languages) pose significant challenges for accurate comprehension and generation. 
  

Programming Languages AI Struggles WithFor coding, AI performance depends heavily on the volume of training data. Languages that are obscure, low-data, or highly specialized are often misunderstood or incorrectly generated:

• Obscure/Niche Languages: Languages like Lisp, HolyC, Racket, D, Erlang, and ELisp are frequently misused because they lack the vast training data available for popular languages.  AI often hallucinates non-existent libraries or mixes syntax from different variants.
  
• Low-Data or Legacy Languages: BASIC variants (especially for 8-bit computers), COBOL, and Assembly are poorly handled due to fragmented training data or lack of modern examples.  AI often fails to distinguish between different BASIC dialects or generates uncompiled C code with memory management errors.
  
• Strongly Typed Systems Languages: While Java, C#, and C++ are common, AI can struggle with their complex abstraction layers and strict typing rules, often producing code that fails to compile or requires extensive human correction. 
  

Languages AI Generally Does Not Use for DevelopmentWhen asking which languages are not generally used by developers for AI projects (rather than which AI struggles to write), the following are typically avoided due to lack of libraries and community support:

• Perl, Fortran, and COBOL: These lack modern AI/ML libraries and are not well-suited for current data science workflows. 
  
• Visual Basic: Primarily associated with legacy desktop applications rather than modern AI development. 
  
• HTML/CSS: These are markup and styling languages, not computational programming languages, and cannot build machine learning models." (LLM)
  
  ----
  
  Using aboriginal languages, mixed together, could confound AI, perhaps.
  

There are cracks there.

[DBCowboy]

I have been fortunate enough to have lived in times when freedom, the Constitution, pride in nationality, actually meant something.

But our country is in the middle of a civil war.  Not one fought primarily with bullets, but with abuses to our legal system.

We actually have a political party now that openly wants to stack the courts, Supreme and otherwise, open our nation's borders, not obey the laws written, add states like DC to alter the make-up of the senate, and gerrymander to maintain a one-party rule for ever.

They say we are not united because so many won't obey their ideology.

The promise punishment and retribution if ever given the opportunity.

Now many will call the opposition the same or even worse.  


I guess, if you're looking for sides in this new civil war, I'd side with the group that will fight for the Constitution, fight for freedom, fight for small government, fight for low er taxes, fight for less spending.

[BeyondKnowledge]

Last month or so I read about a phishing method made to mess up anti-phishing AI agents. It used a specific, specially made, font, in which an A, for example, would look like an M. As AI reads the page's HTML, it read what was really there, with As being As, while humans would read a different thing.

What you are describing is a simple substitution code. Machines read the text as a string of numbers. You can have them show any other letter I  the screen for any number. AI would figure that out quickly.

It would take something more sophisticated. But even the enigma machine was broken by machine under human control.

[ArMaP]

What you are describing is a simple substitution code. Machines read the text as a string of numbers. You can have them show any other letter I  the screen for any number. AI would figure that out quickly.

It would take something more sophisticated. But even the enigma machine was broken by machine under human control.

AI doesn't read it after being rendered on the browser, they would need an image for that, they read the HTML code.
That's why it works.

[BeyondKnowledge]

AI doesn't read it after being rendered on the browser, they would need an image for that, they read the HTML code.
That's why it works.

AI reads it before it gets to the browser. No need to read the screen if you have the instructions on how to make the text.

[ArMaP]

AI reads it before it gets to the browser. No need to read the screen if you have the instructions on how to make the text.

Sorry, I don't understand what you mean by that.

[ReturnofBroccoli]

AI reads it before it gets to the browser. No need to read the screen if you have the instructions on how to make the text.

Traditional AI email filters look at raw data strings (like ASCII/Unicode tokens) to understand intent and catch phishing flags. If an attacker uses font code manipulation to split a word or substitutes standard letters with look-alike foreign glyphs, the NLP model reads gibberish or an entirely benign word. Because the semantic context is broken for the AI, the malicious message passes through security filters undetected.

Advanced AI scanners take screenshots of incoming emails or landing pages and run them through Optical Character Recognition (OCR) to evaluate what a human would see. Because an AI model classifies pixels based on mathematical weights, subtle pixel shifts such as a custom-rendered font merging an "r" and an "n" to form an "m" will trigger a targeted misclassification. The AI reads the visual text as standard text, entirely missing the hidden malicious string underneath