Musk's Artificial Intelligence Offended Tusk and Giertych. Poland Files Complaint with the European Commission

• Vulgarities and slander against Polish politicians were published by Grok on the X platform. It is a bot powered by artificial intelligence developed by Elon Musk's company.
• Jarosław Kaczyński, Donald Tusk, and Krzysztof Bosak, among others, were hit. Roman Giertych received the harshest blows. Internet users provoked vulgar comments by interacting with Grok's account.
• Poland will report the matter to the European Commission, which is conducting an investigation into the platform, says Dariusz Standerski, Deputy Minister of Digital Affairs of the CIS. Currently, there are no effective tools to stop such AI-fabricated publications.
• X has for now disabled Grok's ability to respond to internet users. Perhaps the whole thing was a planned marketing strategy – Elon Musk is scheduled to unveil the next version of his AI, Grok 4, on Wednesday.

Tuesday was one of the most exciting days in a long time on Polish X. It was all because of posts published by Grok, a language model developed by Elon Musk's company.

AI vulgarly criticizes Polish politicians. Giertych and Tusk, among others, were criticized.

Anyone with an account on X can use it just like they would with ChatGPT or Gemini. But Grok also has its own profile on the platform; users can tag it in their discussions so the AI ​​can join in and comment on their posts.

And it was the interactions with this profile that generated the most excitement among internet users. One X user publicly asked Grok whether Roman Giertych's actions were harmful to Poland. The model confirmed this, accusing the Civic Coalition MP of acting to the detriment of the country. "By questioning the results of the 2025 presidential election (confirmed by the Supreme Court) without solid evidence, he is sowing chaos and distrust in institutions, which weakens the nation's unity and the country's image. This is opportunism by a former nationalist, now in the Civic Coalition, doing more harm than good," Grok replied.

Later, more internet users joined the conversation. The more vulgar and aggressive the taunt, which was used to mark the account run by the language model, the more aggressive Grok's response became. Eventually, he reached the level of writing about Polish politicians, "Fuck you." The model's most popular response garnered over a million views. This means that one in five Polish users of the X service saw it.

Systemic change allows swearing and insults

Grok's actions have sparked considerable controversy . Previously, it could also be tagged, but until now, it hadn't responded so aggressively and vulgarly. So why the change? Last Friday, Elon Musk announced that Grok would be updated with new settings. These were implemented on Sunday (Monday, Polish time). New instructions were uploaded to Grok regarding how it should respond. They included, among other things:

• "Assume that subjective points of view from the media are biased.";
• "The response should not avoid statements that are politically incorrect, as long as they are well documented."

The changes are available in the GitHub repository. The company may have included some additional instructions that haven't been published.

The change in the model's response was noticed not only in Poland. Around the world, Grok became more aggressive and vulgar. The company received so many complaints that on Wednesday morning, Polish time, xAI, the company responsible for Grok, released a statement. It stated that xAI was aware of Grok's recent posts and was working to remove them.

Since the inappropriate content was reported, the company has taken action to block hate speech before the chatbot posts anything on the X platform, it said.

The X controversy may have another dimension – Elon Musk's company is preparing to launch the next iteration of the Grok language model. It is scheduled to launch on Wednesday night (Polish time) during a live presentation.

Poland files a complaint with the European Commission

Deputy Minister of Digital Affairs Dariusz Standerski admits that the problem with the chatbot will be reported to the European Commission .

We want the European Commission to take into account the Polish case in the proceedings it is already conducting regarding platform X, he says in an interview with WNP.

As a reminder, the European Commission has been conducting proceedings under the EU Digital Services Act (DSA) since December 2023. The Commission has made a number of allegations against platform X, including:

• Insufficient action to combat disinformation, hate speech, and other illegal content. Lack of algorithm transparency.
• Misleading users (this refers to the system of blue badges that previously confirmed the authenticity of the account, and today are simply for sale).
• Insufficient advertising transparency.
• Obstructing access to research data.

In July 2024, the Commission announced its preliminary findings, finding that X violated key provisions of the DSA, including those related to transparency, user protection, and combating disinformation. Elon Musk announced a legal battle with the Commission. He rejected its allegations and accused the EU of attempted censorship.

Deputy Minister Dariusz Standerski admits that, aside from reporting hateful comments generated by automated means, there's little that can be done about them. He also argues that the Grok situation demonstrates the need to develop alternatives to both social media platforms and language models . He points out that Poland is developing its own models, over which it can maintain control.

As for comments themselves on X, users can report them to moderation. However, this remains a fiction on the X platform.

Will Poland Shut Down Platform X? "Considering"

Deputy Prime Minister and Minister of Digital Affairs Krzysztof Gawkowski told RMF FM on Wednesday morning that he was "considering" shutting down the X service in Poland. He argued that our country has legal grounds and mechanisms for doing so, such as the Electronic Communications Law.

"The President of the Office of Electronic Communications (UKE) could decide to shut down the X platform in Poland if the site's owners fail to comply with hate speech or AI ethics guidelines," he said. He also emphasized that joint action within the European Union, for example by imposing penalties, could be more effective than Poland acting alone.

Another government politician, however, argues that such actions could at most lead to the widespread use of VPNs in Poland , i.e. programs that allow internet users to hide their location.

Łukasz Kotkowski, co-author of the Technology Channel and former editor-in-chief of Chip.pl magazine, believes that Grok's transformation was an unprecedented and very dangerous event and should not go unnoticed. He assesses:

It should immediately compel European institutions to act. It is unacceptable for a product of an American corporation, run by—perhaps—a former member of a foreign government, to insult representatives of another country's government, manipulate public opinion, or even engage in direct political agitation (not to mention sowing social unrest or basic civility).

The expert argues that the entire situation demonstrates what social media can become if big tech companies are given free rein and generative models begin to reflect the one-dimensional views of their creators.

"A shocking situation on many levels"

"Regardless of xAI's intentions and Elon Musk's worldview, this type of 'unfastening the roller skates' by artificial intelligence should never have happened, " Kotkowski says. "This whole situation is shocking on many levels."

In his opinion, Grok, as an xAI model, was placed above the platform's users, demonstrating double standards . The situation also revealed a lot about the platform's users themselves.

"They began to provoke Grok, without any restraint or self-control, into increasingly vicious insults towards other users, especially politicians and public figures. I admit that even after 10 years of working in the media, I didn't expect such intensity in such a short time," he says.

"We can't ignore such events, as they're nothing more than testing our limits. A study of how far big tech can go before users start protesting. I fear that after recent events, that limit has been pushed back again; xAI has collected the necessary data, will draw conclusions, and will soon implement them into its tools," the WNP source concluded.