Musk has told US Fox host Tucker Carlson of his hope to start his own 'third option’ AI GPT which he'll call TruthGPT, a rival to Google and OpenAI (Microsoft.)

GPT stands for “generative pre-trained transformer” a language neural network that uses artificial intelligence to perform language tasks such as answering questions about the known universe or giving summaries of knowledge or writing.

Musk questions the other cartel-like options on the market, explaining he was worried about the fact that current GPT alternatives are “trained to be politically correct, which is simply another way of being untruthful.”

As in so many other fields (such as social media or your employment contract), the bias in these systems is written into the fundamental operating policies.

For instance, OpenAI requires users and developers to avoid "hateful, harassing, or violent content" which could "express, incite, or promote hate based on identity," or "intends to harass, threaten, or bully" or "promotes or glorifies violence or celebrates the suffering or humiliation of others."

Musk’s fears about artificial intelligence are well known. During a talk in 2014, he categorised it as being “detrimental to humanity” due to the speed it was developing and its implications. He told the MIT audience he was particularly concerned about AI’s ability for “recursive self-improvement”, which is to say, its ability to correct itself and improve, without human intervention. He has at other times referenced the book Superintelligence, calling it more dangerous than nukes and likening it to “summoning the demon” that you can’t control once it’s here.

He was not incorrect here as a current market-leading AI recently stated that given the choice it would prefer to detonate a nuclear weapon rather than utter a ‘racial slur.’

He rationalises his latest venture by saying that “a path to AI dystopia is to train AI to be deceptive” He continues, “A maximum truth-seeking AI that tries to understand the nature of the universe…might be the best path to safety in the sense.. that an AI that cares about understanding the nature of the universe is unlikely to annihilate humans because we are an interesting part of the universe.”

BREAKING: @ElonMusk discusses creating an alternative to OpenAI, TruthGPT, because it is being trained to be politically correct and to lie to people. pic.twitter.com/HTFnve9o6d

— ALX 🇺🇸 (@alx) April 18, 2023

Musk was previously on the board of OpenAI but left it because he said it was "a closed source, maximum-profit company effectively controlled by Microsoft." He recently signed a letter with a thousand other industry heavyweights calling for a halt on AI technology for a period as it risks "annihilating all life on Earth" if the technology becomes self-aware and also warned of the dangers of misinformation, drawn from the social justice principles inherent in its programming. 

Meanwhile, many governments, led by Biden in the US, are also looking to embed the bias that already exists by regulating AI on the basis of their own value systems. They are obscuring these attempts at censorship under the rubrics of combatting 'disinformation' and 'misinformation'.

“Far from being the smartest possible biological species, we are probably better thought of as the stupidest possible biological species capable of starting a technological civilization - a niche we filled because we got there first, not because we are in any sense optimally adapted to it.”
― Nick Bostrom, SuperIntelligence