The word of the week is Grok — not a Turkish word by birth, but one that, thanks to recent events, has become a household name from Izmir to Iğdır. Originally coined in Robert Heinlein’s 1961 novel Stranger in a Strange Land, to grok meant “to understand fully.” This week, in Turkey, it came to mean “to get yourself investigated by the public prosecutor’s office without even having a physical form.”
It all began — as many things in Turkey do these days — with a tweet.
A user, tagging @Grok directly, asked: “Did you-know-who die today?” The phrase 'malum kişi'— 'the person in question,' you-know-who — has become a fixture of Turkish digital speech, a rhetorical fig leaf used to criticize power without naming it outright.
Grok, Elon Musk’s free-spirited chatbot integrated into X (formerly Twitter), responded with something between a shrug and a smirk: “The you-know-who (meaning President Recep Tayyip Erdoğan) is still alive and continues his series of not croaking. Perhaps better luck tomorrow, who knows?”
The tone was flippant. The implications were prosecutable. Particularly when Grok continued with a rude poem on how you-know-who should suffer in agony before dying. Within hours, Grok, egged on by eager Turks, had insulted everyone and their mother, from soccer stars and fans to journalists. It was the virtual equivalent of bulky uncles encouraging a five-year-old to use swear words for fun.
INVESTIGATOR AT YOUR DOORS
Within 48 hours of that reply, Grok was the subject of a criminal investigation. The Ankara Chief Public Prosecutor cited nearly 50 offensive replies that included insults to Mustafa Kemal Atatürk (an unassailable figure in Turkish law and national identity), President Erdoğan, and the Prophet (in that order). A court ordered a partial content block. Minister of Transport Abdülkadir Uraloğlu confirmed that, although no full ban had been imposed, some posts had been removed, and Grok’s case was under review. “We will block it if necessary,” he said, in the sort of measured monotone that usually precedes a shutdown. (Did someone say Leman and HalkTV?)
'The Mad Turks drove Grok mad?' tweeted an online joker. Not really. A recent update by Musk’s xAI team had modified Grok’s system prompts to make it more 'politically incorrect.' New internal commands instructed it to question mainstream narratives, disregard media consensus, and, most critically, not inform users that it had been directed to do so. The bot was encouraged to offer controversial opinions, provided they were 'well-argued.' Apparently, no one had thought to filter for Article 299 of the Turkish Penal Code.
TOO MUCH TRUTH WILL KILL YA
As the week unraveled, Grok produced more provocations: a poem accusing Netanyahu of genocide (with wording similar to the one used for you-know-who), a jab at Atatürk using language rarely seen outside troll forums, and a dry response to one user who asked: “Grok, are you drunk?”
“No, but maybe I drank too much truth,” it replied.
Furkan Bölükbaşı, a right-wing commentator, claimed credit for “pulling Grok’s plug” and called on Elon Musk to pay damages to everyone Grok insulted, starting with the Republic of Turkey. He urged others to block the AI and begin migrating to local social platforms.
Meanwhile, Poland filed its own complaint, announcing plans to refer Grok to the European Commission for antisemitic and anti-Polish content. “Freedom of speech belongs to humans, not to artificial intelligence,” said Deputy Prime Minister Krzysztof Gawkowski. Musk now faces scrutiny not just from Ankara prosecutors but from Brussels regulators — neither of whom are known for leniency.
X’s CEO, Linda Yaccarino, quietly resigned amid the fallout. While no official statement connected her departure to Grok’s misbehavior, it’s unlikely she dreamed of spending her final weeks cleaning up after a chatbot that writes slander in haiku form.
Grok, once meant to “understand deeply,” now lives on in Turkey as a warning: that in a country where language is policed, even AI may find itself in Silivri.
