Elon Musk builds his own AI and gives an interview to Tucker Carlson

TruthGPT
Elon Musk wants to build a truth AI – and announces it to Trump fan Tucker Carlson of all people

Elon Musk repeatedly warned of the dangers of artificial intelligence. He also sees the hype AI ChatGPT as a concrete danger. Now he has announced a counter-project.

The tech billionaire has repeatedly emphasized that Elon Musk sees artificial intelligence as a major risk – even recently. In an interview with right-wing US talk show host Tucker Carlson, the co-founder of ChatGPT developer OpenAI now explains why he sees the chatbot as a specific threat. And announces his counter-program.

Musk reported in the interview broadcast on Fox News that he was already dealing with the dangers of AI during his studies. The argument is perfectly understandable. According to Muks, humans have prevailed in nature primarily because they are more intelligent than other creatures. “What if an AI now becomes infinitely more intelligent than the most intelligent human being?” he explains his concern. “The answer is: We have no way of knowing what’s going to happen,” he summarizes. “It has the potential to destroy civilization.”

Fight against the AI ​​downfall

This concern also drove him to co-found OpenAI. They wanted to develop an ethical, transparent AI. Musk cites a conversation with Larry Page, the founder of Google, the world leader in AI at the time, as motivation. He simply didn’t take the danger seriously enough. Now Musk sees the same danger in OpenAI, which he funded with $100 million. “It’s ironic,” he admits.

Together with numerous other experts, Musk therefore signed an open letter in March calling for more government regulations for AI. “I’ve long been a proponent of government regulation in this area, even if it’s not fun to be regulated.” However, as a rule, regulation only comes about when something bad has already happened. “If something happens with AI, it could already be too late. The AI ​​could already have taken control,” Musk fears. When Carlson asks, irritated, whether he got that right, Musk reaffirms his fears: “Absolutely. That’s definitely the direction things are going in.”

Is AI too woke?

Aside from the end of humanity, the billionaire fears another danger. “If you have an AI that writes very well and knows very well what resonates with the public – and is constantly improving it. Now if this AI starts to influence us – how would we even know?” Musk asks the perplexed Carlson. Above all, restrictions on ChatGPT, which are intended to prevent systematic discrimination against minorities, are a thorn in Musk’s side. “They want it to be politically correct,” he explains. He sees that very critically. “They train the AI ​​to lie to us. They teach it not to give us what’s in the data.”

Musk is referring to efforts to get a handle on racial slurs and other types of potentially controversial statements made by AI. The chatbots are therefore cautious on certain topics, sometimes avoiding statements entirely. One reason for this: the AI ​​programs have no moral compass, they actually only report what they have saved from their vast amounts of learning material. In doing so, however, they reproduce social differences. If a job is mainly practiced by men, the AI ​​could conclude that women are not suitable for it. The developers try to prevent that.

AI as mind control

Carlson draws a radical conclusion from this: “The government wants to control how AI is developed at an early stage. In order to be able to build in dishonesty. This is how the thinking of the American people is to be controlled,” he says. “This is a direct attack on democracy!”

Unsurprisingly, the moderator, who is politically far right, uses the interview to advance his own agenda. The current democratic US government uses technology to manipulate citizens, he claims. “If you go into the app and ask about the government’s crime statistics – they will lie to you,” he claims on “Fox and Friends” when asked about the conversation via ChatGPT. And repeats the well-known accusation from the right-wing camp: the AI ​​is too woke, would censor out of consideration for supposed sensitivities.

In fact, that’s not even true when it comes to crime rates in the US: If you ask ChatGPT which social group commits the most crimes there, the bot reports that black men were the most frequently arrested for murder in 2015 – but at the same time points out that these statistics alone do not provide a complete picture.

Elon Musk: Truth Saviors the World

But for Musk, it’s about more than just freedom of speech. He uses the conversation to announce his own AI. “We’re running late,” he admits. But he still wants to try. His project TruthGPT is “an AI that wants to reach the highest level of truth and understand the nature of the universe,” he explains. “That could be the main route to safety. If an AI wants to understand the universe, it’s unlikely that it will wipe out humanity. Because we’re an interesting part of the universe,” Musk said. “We could also just go and kill all the chimpanzees. But we’re not doing that.”

Source: FoxNews