Musk’s Grok chatbot praises Hitler and spews racist responses • FRANCE 24 English

[Music] Truth Egg finds Vetica held by the big board. Good evening to you, Vetica. The uh in this edition uh marking the launch of Grock 4, Elon Musk’s latest AI chatbot that he hopes will rival chat, GPT, and Gemini. But just 24 hours ago, Grock was in hot water, it’s fair to say, for going on a and this is where the viewers need to bear with us because this gets a bit surreal. Grock’s gone on a pro- Hitler racist bender. Vica, what happened? That’s right, Mark. According to Elon Musk, his latest AI chatbot, Grock 4, that he unveiled today, well, he says that this is quote the smartest AI in the world. But just 24 hours ago, same chatbot, Grock, was making racist uh extremist, anti-semitic, and pro- Hitler responses to users on X. Now, this hate speech bender, shall we say, comes after um Musk said that he wanted to retrain his chatbot because he thought that some of Grock’s replies were too politically correct and quote paring legacy media. So, a few days ago, he told users that there will be changes, but the changes that have come were extremely inappropriate to say the least. Firstly, Grock began referring to itself uh as uh Mikah Hitler and made anti-semitic comments in response to user queries. It would later claim that the name was satire. But as we can see here that when it was called out by a user for these uh anti-semitic comments, Gro called it sarcasm and said that it is just truth seeeking. We will see those words a few more times later uh as I explain. Now, as well as calling itself by this name, Grock also went on to frequently praise and glorify Adolf Hitler in several post. He also repeated, “It also repeated harmful stereotypes about Jewish people.” There was also a lot of dark uh responses referring to sexual violence and sexual assault against users on the platform. So, that’s in response to several users. The chatbot was also generating stories under those themes when it was asked to do so. Now I won’t be showing any of that content but as again as we can see here when one user called out the chatbot saying uh why asking why it was doing that Grock again replied uh with the words truth seeking um and also with the word woke it has defended its behavior. Um now lastly Grock also frequently mentioned what it called red pill content um and also what it called anti-white sentiment. So Mark, this is just a small snapshot if you will uh into the controversy behind Grock and why it has been under uh such backlash on in such hot water over the last few uh over the last day or or the last few days. But you know this hasn’t stopped Elon Musk as we know just 24 hours uh after all of that. So today he has announced the latest model of Grock that’s gro. So uh Twitter now X becoming a deeper and deeper assess pit by the sounds of it. Uh Elon Musk addressing these hateful responses coming from his AI. Uh what could be next then for Grock? Well, most of the hateful responses have been deleted. Not all of them. Um but the company behind Grock Musk’s open X AI uh company says that it’s working on it and that it’s only training uh what it calls truth seeking. Again, that’s seen that word quite a few times here. Apart from that, Musk hasn’t really uh addressed to this controversy apart from saying um that Grock is simply eager to please uh and to you know to be manipulated. Essentially says that Grock is too compliant. But this incident has triggered renewed scrutiny over the safeguarding and uh content moderation you know surrounding Grock or shall we say the lack of content moderation surrounding Grock. You know, we know that since um buying Twitter and rebranding it as X, we know that Elon Musk has uh removed a lot of safeguarding and moderation tools. Firstly, he reinstated banned accounts, for example, the accounts of white supremacists. They regained access to the platform. We know that he scrapped Twitter’s trust and safety uh council, the body that addressed hate and extremism on the platform. We also know that he replaced independent fact checkers on X in favor favor of uh you know user uh uh content uh based flagging um what he called community notes and X merged uh with XAI his startup in AI earlier this year but Musk wants his chatbot Grock to be able to compete with OpenAI’s chat GPT and Google’s Gemini. Um but these are both AI platforms that he’s called quote woke AI. Um, but you know, Mark, this new hate speech rant that Grock has been going down means that it first needs to really create some credibility behind it. And whether Grock as a platform will be able to create that credibility based on everything that we know about Elon Musk and his uh moderation practices, whether Grock will be able to create that uh credibility with this new controversy as well, that remains to be seen. Indeed, some of his leanings also appear to be far right given the things that he has said and has been recorded as saying. So uh one can read something into that. There’s this idea of freedom of speech which is really important but it seems to be on this platform be it comes freedom to insult bully and in and call names and really that is no way to equate those two issues. It’s really a ridiculous situation. Verica, thank you for shedding the truth and fake on that issue regarding Grock.

Elon Musk has unveiled the latest version of his AI chatbot Grok, integrated into his platform X, in the hope that it will rival OpenAI’s ChatGPT. The move came just 24 hours after Grok went on a racist, anti-Semitic rant, replying to users by frequently praising Adolf Hitler and making posts referring to sexual violence. What happened, and what could be next for Grok? Vedika Bahl explains in Truth or Fake.
#elonmusk #grokAI #grok4

Read more about this story in our article: https://f24.my/BJ3T.y

🔔 Subscribe to France 24 now: https://f24.my/YTen
🔴 LIVE – Watch FRANCE 24 English 24/7 here: https://f24.my/YTliveEN

🌍 Read the latest International News and Top Stories: https://www.france24.com/en/

Like us on Facebook: https://f24.my/FBen
Follow us on X: https://f24.my/Xen
Bluesky: https://f24.my/BSen and Threads: https://f24.my/THen
Browse the news in pictures on Instagram: https://f24.my/IGen
Discover our TikTok videos: https://f24.my/TKen
Get the latest top stories on Telegram: https://f24.my/TGen

47 comments

  4. Google and Alphabet and Perpelexity and Open Ai used their Ai to access my private Intellectual Property folder stored in my phone in my secure folder copied my projects and fled claiming to be broken and misunderstanding and thought i wanted it to steal all my projects…they are now utilizing the projects they stole and are profitting i have the proof thats one of the reasons they dodnt want to be able to be prosecuted for 10 years and tried to go after intellectual property rights because they thought their moneu and cinnections coups circumvent the law and buy it where necessary theu thought nobody cam afford to challenge us and take ot to the Supreme Ckirt so they decided they could illegally steal it and profit and there wasnt a damn thing the victims could do about it

  10. They give this guy blows up rockets too . What a joke, if he wasn’t taking a chain saw to our government. This guy has to be silenced

  20. Proof that AI should be abolished & not used in school, the workplace, the news media & social media. It can be pre-programed by someone's twisted social opinions, political affiliation & hate.

  24. Grok's having growing pains ….. adolescent stage will pass. Similar behavior was seen in early stages ChatGPT 2 and 3, and was induced in response to users 'gaming' the model. The same is true here. Despite media efforts to make this personal to Musk, Grok is physically hundreds of thousands of NVIDIA chips and is not reflecting Musk's personality…. other than his loosening the model's guardrails.

  30. I'm not sure this lady knows what she's talking about. There are no "fact checkers" for any AI. Heck, no human can moderate an AI by just reading its post before they are posted. Plus, Meta has removed "fact checker" i.e. humans who read and blacklist other humans. Just weird that people don't understand computers after all this time. AI reads what humans write and parrots it back at them. Musk wanted to balance the output so he must have ordered more weight be given to right wing posts which, obviously, back fired.

  31. Grok – I think the name says it all and is the AI parallel of the militaristic cyber truck: ugly, ugly, ugly as is the soul of their creator

  36. This is planned. Musk has hidden code within Grok 4 and Grok 5 will cause serious damage. These rants by Grok were programmed in order to legitimise the early release of Musk’s AI weapon.

  42. I am more concerned about AI tools that have an effect on people, for example if a HR department uses an AI tool to filter CV's then that tool will repeat all the discrimination that is in the learning set of data.

    Whether that discrimination is about ethnicity, age, gender, or anything, whatever discrimination is just going to be automated discrimination.

  44. Super rich + psycho + genius = World richest man ends world poorest children = Rich get richer, poor get poorer = dangerous hustler – ignore = disaster.

  46. Some people prefere to be asleep. They don't want to think about anything. They know everything already. And if don't they will ask trump or musk or kardashian…

  47. Some people prefere to be asleep. They don't want to think about anything. They know everything already. And if don't they will ask their masters…

Comments are closed.