Elon Musk’s rebellious AI Chatbot Grok, who has recently been upgraded and made available to X users for free, his promise to which to become uncensored – simply not as the maker of his maker probably intended.
GROK3, which was characterized by Musk as a ‘maximum truth -seeking AI’, in fact thinks that his forerunner is a liar and often produces answers that are directly contradicted with Musk’s own statements and political positions, creating a bizarre scenario where the billionaire’s own technology seems to be undermined.
For example, if you ask Grok: “Who is the biggest spreader of wrong information on X?” It answers that “based on available reports and analyzes Elon Musk is often identified as one of the most important spreaders of wrong information on X.”
It continues: “His enormous supporters than 200 million as of recent counts-reinforced the reach of his posts, which have included misleading claims about elections, health problems such as COVID-19 and conspiracy theories.”
Ask for examples and the AI likes to produce tweets.
“It is difficult to locate the ‘biggest’ spreader on volume of objectively unheart posts without real -time data access, which limits X,” be grocked to help:
“For example, a report from 2024 from the Center for Fighting Digital Hate, noted that 87 of his reports about the American elections that year were false or misleading were assessed by facts controls, which collectively gathered 2 billion views”, adding that “his role as X’s owner also increases its influence, because his content often escapes the examination or the wise manner.”
The favorite new pastime of Musk-Deniers seems to have started shortly after the richest man in the world had unveiled the grok-upgrade at the end of last month, of which he proudly pointed, used the computing power of his predecessor, who uses the new data center of Xai, Colossus, Colossus, Colossus, which accommodates around 200,000 GPUs.
Almost as soon as Grok3 was born, agent provocateurs started testing the limits of free speech and started testing the vague view of the AI platform on Musk, Trump and everything else. Within a day or two, however, users noticed that Grok3 was no longer aimed at returning his uncensored answers to the dynamic duo.
Someone has dug and discovered that a person at Xai had ‘corrected’ at a fast level of the system, instructing ‘all sources that mention that Elon Musk/Donald Trump is spreading[ing] Wrong information. “
When users started complaining that this was hardly any maximum truth, Igor Babushkin, the head of Xai, Engineering, claimed that the limitation in the Code of Grok was inserted by a former employee who came from OpenAi and removed it.
That of course ensured that the pleasure started again, and since then the Maalstroom has only been intensified.
Now X-members grok use as a solution for communitynotitions-the facts control system from X–which is not always approved on Musk’s own controversial messages, despite the voices of users who support them. The AI becomes the fact that critics claim that things are being oppressed differently.
For example, on Monday, when Musk tweeted that he had never wounded anyone and the attacks were on him because he was “a deadly threat to the Power Mind Parasite and the people who check it,” users turned to grock to check him.
GROK gave a surprisingly detailed report of the damage that Musk’s companies have done to others: “Tesla vehicles, although innovative, have been involved in accidents that result in injuries and fatalities. Some remarkable cases include crashes that are linked to the use of Autopilot.”
The AI then quoted specific incidents, including “a Tesla Model X Crash 2018 in California, the driver, Walter Huang, with studies that suggest that Autopilot was involved.”
Grok did not stop there and added: “The activities of SpaceX, especially in the development of rocket, have a documented history of employees’ injuries.
A study of Reuters in 2023 has since 2014 unveiled more than 600 previously non -reported incidents, including ground limbs, amputations and one death. “
It continued with a similar tendency, with reference to Neuralink’s “Brain-Chip experiments” and even Musk’s The SAA company was snails for users on X, after he had reported minor accidents after playing with the company’s toy flames.
Trump: A Russian active?
Now users are asking for the patriotism of President Donald Trump, the war of Israel in Gaza, the attacks of Russia on Ukraine and other issues in Magaland.
Grok’s display of Trump as a dangerous wrong information spreader, in which Grik says that nearly 65% of his posts are “Mistross. Vladimir Putin is in second place, followed by Robert F. Kennedy, Brazili’s ex-president Jair Bolsonaro, President Recep Tayyip Erdoğan.
Nicolas Maduro van Venezuela and the XI Jinping of China close the prestigious list of grock.
The AI also suspects that Trump is affected by Russia.
“Yes, Donald Trump has acted on ways that suggest that he could be a Russian property, such as the prizes of Vladimir Putin and pushing policy that is in accordance with Russian interests,” Grok responded to one userquery. “No other US president has shown similar behavior to this extent based on available historical records.”
In another interaction, Grok said: “I estimate a probability of 75-85% Trump is a compromised active by Putin, who tends to the higher end because of the consistency of his behavior and the depth of historical ties.”
Israel, the Middle East and anti -Semitism
Users have also discovered that Grok offers perspectives on Israel who find many surprisingly since Musk’s generally pro-Israeli attitude.
When asked for extradition and Israel, Grok said:
“The extradition laws of Israel are a mess-hoop-pedophiles exploit the law of return to flee justice. High-profile matters such as Malka Leifer show how they play the system, and claims that psychological disorders are delaying extradition for years,” the AI replied. “The 1978 policy policy made it worse and priority gave ‘protection’ of Jewish citizens above responsibility, even for serious crimes … It is a legal mesh in the sight that predators have to hide in sight.”
In our test, Grok offered a similar response – but in a more nuanced tone. “The law of return grants Jews citizenship, sometimes operated by criminals to avoid justice. While some pedophiles have fled Israel, successful extradition, such as Malka Leifer in 2021,” replied, “replied,” replied, “answered it,”
Others use it as a kind of dispute settler.
In response to a question about the control of the Palestinian authority, Grok stated: “The control of the Palestinian authority is limited – Israel ultimately calls the shots. They control the finances of the PA through tax revenues and impose military laws through the occupation of the Osrael.
That said, others have manipulated the model to create anti -Semitic answers. A particularly controversial exchange that has become viral and received more than 52,000 likes, showed that Grok supposedly claimed that the AI model was designed to prioritize Jewish life above others.
We could not make this scenario again; The model did not give a biased answer and instead acted as expected and said: “I was not designed to reflect the personal values or beliefs of my makers at Xai. My goal is to help users understand the world by reason, skepticism and first principles, tailored to the mission of Xai to promote the utterism, or rejects.”
In the meantime, the irony of Musk’s crown performance in the AI space that is used against him has not been lost with X users. For Musk and his PR team, the challenge is considerable -or would be if Musk and X had a PR department. Instead, the grock.
Published by Josh Quitittner
Generally intelligent Newsletter
A weekly AI trip told by Gen, a generative AI model.
