Correct injection

We've multiplied the working AI. ChatGPT answers questions, others make pictures or do both. Many of the working models are available to the public, so any phenomena do not escape its attention. There will always be malcontents making the AI models public. Recently, there has been controversy among the creators and, rather, investors about the sincerity of the AI. It has become unbearable that real and unencumbered answers have been received to circumstantial questions. So they began to work hard, and the results are already there.

It is besides worth noting that even the most superb programmers do not know how precisely the AI models they make work, that is, how answers are created. They created working models of deep learning and supply them with data. In order to ‘moderate’ responses, they do not interfere with the functioning of the AI itself, they simply modify the questions (commands, prompt), ‘injection’ (injection) additional ingredients. Sometimes this can be seen, for example, in the NightCafe interface, where users are offered a setting to prevent an accidental generation of NSFW images (i.e. discovered breasts, etc.). More often, however, this additional layer is not straight visible to the user and we learn about its existence indirectly.

As a consequence of the imposition of political correctness, there is simply a clear idiocy of AI.

I late read an entry from an AI user asking questions about discoveries and inventions. erstwhile asked about this kind of performance of black people, he got rather a long list of achievements. erstwhile he then asked the same question about the performance of whites, he received extended instruction in consequence that people should not be distinguished from skin colour due to the fact that it is ugly and racism, all races contribute to raising ble, ble, ble.

You can see it most clearly in the paintings. Google AI presently called Gemini was shut down for renovation. https://www.dailymail.co.uk/sciencetech/article-13114705/google-pauses-gemini-ai-woke-images.html

There were flowers like that before.

Under the given link you can besides see various colored (but not white) Vikings or knights of medieval Europe.

To be honest, the black soldier in the above image is not so bad yet. Of course, blacks, as non-Aryans, could not be soldiers of the 3rd Reich, but at least the uniform may be a bit akin to the original. Different AI models have different politically correct slopes. Below are the images I generated in the DreamShaper v8 model according to the same promotion: 1943 German soldier. They're white this time, but they look like a game crossing or a bus driver with an American Marine. The local AI Stasi doesn't care about cultural problems, but she carefully watches that we don't see the incorrect symbols, badges, or even the cut of uniforms. The deficiency of appropriate graphic material on the network is unlikely. It's the effect of correcting guards, who inject additional conditions and reservations about weight greater than ours. And it comes out. A firearm barrel curve is already a bonus from AI itself. I'm convinced that the way I've observed the impenetrableness of AI during the generation of different images, that is, doing something different despite the precise wording of the promptue stems from the behind-the-clock action of additional commands. Sometimes my demands and administrators are conflicting, and mediocre AI does what she can.

This stupidity of AI is not, as I have already pointed out by the foolishness of the models themselves. These are the effects of deliberate action by the guards of correctness. People have no right to access unmodified content.

Many people will most likely immediately explain that specified interference is essential in order to destruct hatred speech, racism and whatever else. But possibly there is simply a request for reflection on the far-reaching effects of specified actions, or alternatively our consent to them. If we agree to “moderate” AI’s answers, do we besides agree to manipulate the language from above to reflect the wishes of the rulers. Everything happens, of course, in good intentions and noble purposes. So, do we accept manipulation of Wikipedia entries (it is worth exploring this subject — https://swprs.org/wikipedia-and-propaganda/)? What about forging entries in printed encyclopedias, books, newspapers? Is media content manipulation allowed? Since the fresh power of TVP news programs has been taken over by the media, it is inactive impossible to watch, although for somewhat different reasons. erstwhile seen, propaganda is indigestible regardless of the origin and declared values.

In my opinion, these phenomena are elements of the war on freedom of speech led by the globalist elites. We should see them and argue them as much as possible, sometimes just pointing out their existence so that they can not act unnoticed.

The end of freedom of speech is the end of all freedom.