It is speculated that ChatGPT may replace many professions: teacher, journalist, legal assistant, anyone who writes for a living.
ChatGPT scours the Web and internalizes the majority beliefs about everything it encounters. Thus the recent buzz about how ChatGPT declared it would be better to allow a nuke to slay millions than to disarm it by saying out loud the Dreaded N-Word.
It’s possible that ChatGPT is being controlled so that it does not simply internalize the majority view but has a baked-in woke bias. For this article I’m assuming it is as presented.
Since it only “knows” the majority beliefs (or rather, the textual expressions of these beliefs) which exist on the Web, if ChatGPT takes over the bulk of mankind’s production of text then heterodoxy and heresy will be averaged out. Outlying concepts and beliefs will be averaged away from the contents of the Web as ChatGPT ignores outliers and swamps them with its own output. Outlying beliefs will become more and more rare, and regime-compliant beliefs will become increasingly common.
If its output is posted on “reputable” sites (schools, government bureaus, mainstream journalism, and so on), ChatGPT’s writings will arrive pre-approved for consumption by itself and other AI entities. Under the current system some heresy does receive the regime’s imprimatur, at least until detected and removed. Not any more, if ChatGPT and his brothers take over.
(In using the word “his,” I’m speaking metaphorically. I don’t believe these entities are self-aware.)
And since the Web has become the de-facto determiner of reality for citizens of the more advanced nations, ChatGPT could wash away heterodox thought from all but a tiny minority.
*
The intellectual effect of ChatGPT is fundamentally different from that of Google and the other regime-compliant search engines. These have done much to spread woke orthodoxy and suppress heresy, but the output of a traditional search engine is mediated through a human being who must decide which hits give the best answers to the query, and who must compile the information to create his own document. Not so ChatGPT. It presents the entire document as a finished product direct from the oracle. The human mediator’s only choice is to take it or leave it, but in the dystopia imagined here there would be no reason to reject ChatGPT’s output. AI could drive us toward intellectual heat death.
The Orthosphere 
“AI” is the chimera of all chimeras. Attempting to achieve it is extreme self-idolatry on the part of the people trying to create it, because intelligence is a gift to us from the Creator, and it is not in our power as creatures to give it. People trying to create it are trying to play God. It is appropriate that Apple inc.’s logo is an apple with a bite out of it, just like the apple offered by the snake and Eve in the garden. Just bite ! And the result is guaranteed to be evil.
I agree. ChatGPT and its brethren seem more than a little demonic.
The blog raises pertinent concerns about AI’s potential impact on intellectual diversity. However, I’d argue that mainstream ideas have always overshadowed less popular perspectives, even before the advent of AI.
Generative AI like ChatGPT could indeed offer a new platform for lesser-known perspectives. They have the potential to amplify a variety of voices and ideas, provided they’re trained on a diverse range of data. This might break the cycle of people echoing mainstream opinions, and encourage a broader, more balanced discourse.
This of course does not discount some of the dangers of AI, some of which are existential, and the need to align them quickly.