This is not investment advice. The author has no position in any of the stocks mentioned. Wccftech.com has a disclosure and ethics policy.

After making a number of stealthy yet decisive moves, Elon Musk has finally spilled the proverbial beans on his AI-related ambitions, and they revolve around a “maximum truth-seeking AI,” simply dubbed the TruthGPT.

NEWS: @elonmusk says he is creating a maximum truth seeking AI chatbot called “TruthGPT” that tries to understand the nature of the universe.

“An AI that cares about understanding the universe is unlikely to annihilate humans because we are an interesting part of the universe.” pic.twitter.com/hAeEzThqqK

— Sawyer Merritt (@SawyerMerritt) April 17, 2023

In an exclusive interview with Fox News host Tucker Carlson, Elon Musk declared that he is working on an AI that tries to understand the nature of the universe and is, therefore, unlikely to “annihilate humans.”

This declaration comes after Elon Musk poached talent from DeepMind – a subsidiary of the tech giant Alphabet – and purchased around 10,000 GPUs from NVIDIA to train a Large Language Model (LLM) on the pattern of OpenAI’s ChatGPT. Additionally, the CEO of Tesla and Twitter has registered a company in Nevada by the name X.AI Corp. These developments show intense interest on Musk’s part to corner at least a slice of the booming AI sphere.

Of course, we do not yet know what exactly Elon Musk means when he refers to TruthGPT as a truth-seeking AI. Is he referring to an AI system that is self-aware and capable of discerning truth from falsehood? If so, this would remain a pipe dream for a very long time. First, our current technologies do not yet allow for a truly self-aware AI. What’s more, truths are quite hard to quantify. Who becomes the ultimate arbiter of what constitutes a truth and what is a falsehood as the AI is undergoing training? Prima facie, this appears to be a perfect recipe for introducing severe biases into TruthGPT.

If, on the other hand, given the naming convention that Elon Musk has borrowed from ChatGPT, we assume that TruthGPT is simply going to be a large language model (LLM), some people might think that synergies with Twitter – and the vast troves of internal data that the global town square generates each day – might give the nascent AI a crucial edge. However, using this internal data from Twitter comes with its own host of challenges, as highlighted by ChatGPT in response to our query:

ChatGPT's response on the difficulty of using Twitter's raw internal data to train a large language model (LLM)

Basically, the raw internal data generated by Twitter varies in quality and consistency, which would undermine the training process of an LLM. Cleaning and pre-processing this data alone would consume an exorbitant number of resources.

Regardless of whatever shape the still-fuzzy TruthGPT ends up adopting, Elon Musk, for now at least, appears to be dead serious about pursuing the ongoing AI race.