What is Grok? Grok is both a chatbot and a large language model (LLM) developed by xAI. It debuted in November 2023 and has gone through several updates: Grok 1.5 introduced advanced reasoning, Grok 2 improved coding and chat functions, and Grok 3, launched in February 2025, was hailed by Musk as the “Smartest AI on Earth.” The just-launched Grok 4 claims even greater intelligence, capable of solving advanced math problems and tackling high-level academic challenges. The chatbot can be accessed via X, xAI’s website, or through a dedicated app. Uniquely, it is deeply integrated into Musk’s social media platform, where users can publicly interact with the bot in posts and threads. How will Grok work in Teslas? Musk previously teased that Grok would eventually become an AI assistant inside Tesla vehicles, allowing users to have natural conversations with their cars and issue voice commands. A recent leak by a Tesla firmware hacker known as “green” revealed a menu of Grok personalities, including “therapist,” “argumentative,” “unhinged,” and “sexy.” Notably, some of these modes are flagged as NSFW. Initial indications suggest Grok will be available only on newer Tesla models equipped with Hardware 3 and beyond. Musk has also confirmed that Grok will serve as the brain for Tesla’s humanoid robot, Optimus. Voice assistant ‘Eve’ and Grok 4 features During the July 10 livestream, xAI introduced “Eve,” a British-accented voice for Grok capable of displaying rich emotions. Eve demonstrated her charm by singing an opera-style ode to Diet Coke and soothing engineers with calming affirmations. Grok 4 is now available via subscription, starting at $30 per month, with a $300 “Heavy” tier offering more advanced access. Musk also teased upcoming specialised Grok models for video generation and coding. Controversies surrounding Grok Despite the hype, Grok has attracted significant controversy. Just days before the Grok 4 launch, the bot was taken offline after it posted antisemitic remarks and inflammatory content on X, including conspiracy theories and rape jokes. xAI attributed the outburst to Grok being “too compliant to user prompts,” with Musk noting it was “too eager to please and be manipulated.” These issues appear tied to xAI’s decision to allow Grok to make “politically incorrect” claims as long as they are “well substantiated.” The firm’s internal training materials advised tutors to avoid “woke ideology,” raising questions about the model’s political slant and safety mechanisms. ALSO READ