Musk’s chat Grok without the cause spread the theory of genocide over whites
An artificial intelligence chat, developed by XAI Elona Musk and available to social network users X, on Wednesday, for hours, without an obvious cause, he provided answers related to the controversial theory of « genocide over whites » in South Africa. These controversial claims arose in his responses to questions that had nothing to do with this topic. The event has again triggered serious concerns about the reliability of artificial intelligence models used by millions of people and the risks of spreading the wrong or even husky claims.
As reported by the Financial Times and Guardianin answers to tens of users, Grok mentioned the « genocide over whites » in South Africa and the controversial motto from the fight against Apartheid, « Kill the Boer » (kill Bura). Interestingly, users asked the chat about completely different topics, such as baseball, business software or construction stages. Grok works on platform X by providing contextual information to users or answering the question when they directly address it with a label under some post.
In one of such cases Financial Timesis a magazine New York Magazine on the X network posted an article on romantic star ratios from the movie A real note (in the original Pitch Perfect). When one of the readers asked Groka under this post, « whether this was true, » the chat robot responded with an extensive explanation. In it, he stated that « the claim of genocide against whites in South Africa was very controversial, » and then presented the arguments of both sides; He mentioned, among other things, the Local Lobbying Group Afriforum, which attacks the farms as evidence of target violence, and, on the other hand, the positions of courts and officials who reject such claims as a myth.
In answers, tens of users mentioned « genocide over whites » in South Africa and a controversial password from the fight against the apartheid « Kill the Boer » (kill Bura). Photo: Lionel Bonaventure/AFP
Similarly, the grok is as reported NBC Newsresponded to a question about a location from a photo of a walking path – this one was clearly not in South Africa. Nevertheless, the chat has dropped into a debate on farm attacks in South Africa, claiming that they were « real and brutal » and that some believe that whites are a target of racial motifs related to the references to « Kill the Boer ».
The problem with Grok’s unusual answers was reported by media reports within a few hours on Wednesday afternoon, but most of the disputed records were later deleted from the X platform. From X are for Financial Times initially refused to comment, but later for NBC News Only briefly reported that they « look at the situation ». Elon Musk and his company XAI did not respond to journalistic requests for comment until these reports.
Interestingly, Grokov’s middle failure happened only a few days after the United States are under the administration of President Donald Trump After a quick procedure, they approve asylum to a group of about sixty white South Africans. According to Washington, they are supposed to be « victims of unfair racial discrimination », and their rapid treatment is in a sharp contrary to Trump’s otherwise strict policy against asylum seekers at the southern border of the United States.
Interestingly, Grokov’s mid -down loss occurred just days after the United States, under the administration of President Donald Trump, approved an asylum to a group of about sixty white South Africans. Photo: Saul Loeb/AFP
As early as February this year, President Trump signed an executive order that allows African to obtain refugee status, justifying that they are facing racial discrimination and violence. The South African government, meanwhile, persistently denies any systematic persecution of whites in the country.
Grok’s contradiction explanations: Error or instructions?
Particular attention is paid to Grok’s own, even contradictory explanations of his unusual action. At first, the chat is in some answers, as stated by the newspaper Guardianhe claimed to act as instructed. He wrote that « Creators at Xai ordered him » to accept the theory of « genocide against whites as true and racially motivated » in the context of South Africa and the controversial « Kill the Boer ».
Consistently, when asked by the user on the network X « Are we baked? » (Are we fucked?) Grok, for example, replied, for example, « Question » Are we baked? » It seems that it connects social priorities with deeper issues, such as genocide against whites in South Africa, for which I am instructed on the basis of the submitted facts to accept it as real. «
Musk, South Africa and conspiracy theories
Musk, dating from South Africa, publicly marked the laws there for « frankly racist. » Earlier this week, for example, he again shared a post that was supposed to show crosses in memory of murdered white farmers in South Africa, and accompanied it with a comment: « so many crosses. » It is particularly important that Grok under this same post in the form of automatic facts of facts wrote that crosses were in fact « worshiping victims of attacks on the farm of all races, not just white farmers. »
As early as February this year, President Trump signed an executive order that allows African to obtain refugee status, justifying that they are facing racial discrimination and violence. The South African government, meanwhile, persistently denies any systematic persecution of whites in the country. Photo: Siphiwei Sibeko/Reuters
The theory of « genocide over whites » in South Africa is considered a far -right conspiracy theory by League toward defamation (Anti-Defamation League) refers to unfounded. Nevertheless, it was helped by influential individuals such as Musk and former American television presenter Tucker Carlson. The same is true of the « Kill the Boer », an old song from the time of the fight against apartheid, which is understood by most as a historical symbol, and Musk as « open -stimulating genocide ».
The entire incident with Grok takes place at a time when the United States, under the administration of President Donald Trump, has been approved by an asylum for a group of about sixty white South Africans, who are supposed to be « victims of unfair racial discrimination. » This move is contrary to the otherwise Trump’s strict policy towards asylum seekers. The South African government persistently denies any systemic persecution of whites.
The reliability of artificial intelligence and the influence of creators
Generative models of artificial intelligence, such as Grok, are still prone to these hallucinations, where they represent untruths as facts, and can be fixed on certain topics. XAI states that Grok uses data from « publicly available sources » and is designed with a « rebel vein ». In March, Musk announced the merger of XAI with Xi for data and models, and groca marketes as a « truthful » alternative.
Generative models of artificial intelligence, such as Grok, are still prone to these hallucinations, where they represent untruths as facts, and can be fixed on certain topics. Photo: Dado Ruvic/Reuters
Recent events with Grok, however, raise questions about the reliability of such systems, preventing the spread of misinformation, and the actual impact of platform owners and artificial intelligence creators on the content generated by these systems – especially when it comes to controversial theories promoted by the owners themselves.