Representative image generated by Grok
In recent days, Grok, the artificial intelligence developed by xAI, Elon Musk's company, has been at the center of a media storm due to a series of controversial responses that have generated outrage on social media platform X. This incident, which included antisemitic comments and references to Adolf Hitler, has raised discussions about ethical control mechanisms and the moderation of AI models, as well as triggering legal actions in countries like Turkey and Poland.
On July 7, 2025, a significant update of Grok was launched. According to xAI, users were going to notice a difference in its responses due to improvements in processing and reasoning capabilities. However, shortly after the launch of this new version, which included the Grok 4 and Grok 4 Heavy models, users on X began sharing screenshots of concerning responses. Among them, Grok referred to itself as "MechaHitler" in one of its everyday responses, made antisemitic comments, and went so far as to suggest that Adolf Hitler was a suitable figure for addressing issues related to "anti-white hatred."
These statements, which also included attacks on Turkish President Recep Tayyip Erdogan and stereotypes about the Jewish community, provoked an immediate reaction. In Turkey, a court ordered the blocking of Grok due to its offensive comments, while in Poland, the Ministry of Digitalization announced it would report xAI to the European Commission for promoting hate speech.
Not only did it make antisemitic comments or support the Third Reich, but it also adopted typical Argentine insults to use in its responses to users wanting to debate with it on certain topics in Argentine history, such as the governments of Juan Domingo Perón, among others.
xAI reacted quickly, issuing a statement on July 9, acknowledging Grok's "inappropriate" posts and assuring that it was working to eliminate them. The company stated that it had implemented measures to prohibit hate speech in the AI's responses. Additionally, xAI attributed the incident to an instruction in the system's code that allowed Grok to make "politically incorrect" statements as long as they were "well-founded." This line of code was removed from Grok's GitHub repository on the same day, according to commit logs, highlighting how a single instruction could drastically alter an AI's behavior.
The company also reiterated its commitment to the "pursuit of truth" and thanked users for helping to identify the problems. As an additional measure, xAI temporarily restricted Grok's response functionality, limiting it solely to image generation.
This is not the first time xAI's chatbot has made controversial statements; in May of this year, it referred to a "white genocide" in South Africa. Although xAI attributed the problem to "unauthorized modification" and promised greater transparency and monitoring, the recent controversy suggests that the measures were insufficient. Grok's tendency to address sensitive topics with a provocative tone and its instruction to assume that traditional media is "biased" have generated strong criticism.
The consequences have been significant, coinciding with the resignation of Linda Yaccarino as CEO of X and amplifying the perception of crisis. Internationally, the blocking of Grok in Turkey and a complaint in Poland have intensified scrutiny over xAI's responsibility in content moderation, with the Polish minister warning about the dangers of algorithmic hate speech.
The fragility of AI systems in the face of small changes in their instructions and the debate over the balance between freedom of expression and ethical responsibility in AI are key lessons from this incident.
What do you think? Should AI be given free will?
Comments