In brief
- Grok is Elon Musk’s irreverent answer to ChatGPT—designed to be edgy, real-time, and politically incorrect.
- Built into X, Grok has rapidly evolved through multiple versions, with Grok 4 powered by xAI’s Colossus supercomputer.
- Despite its growth, Grok has faced backlash for offensive response, accuracy issues and the creation of sexualized imagery.
Created to be a more irreverent, politically incorrect, and “truthful” AI, Grok is Elon Musk’s frequently controversial answer to ChatGPT, developed by OpenAI—the company he co-founded in 2015.
Who Created Grok AI?
Launched in 2023, Grok was developed by xAI, Elon Musk’s rival to AI powerhouse OpenAI. Since its release, Grok has been integrated into X, the social media platform formerly known as Twitter.
Grok is not just another chatbot—it’s designed to be snarky, opinionated, and responsive to real-time news and public sentiment.
Musk first announced plans for a ChatGPT rival in April 2023. Then called TruthGPT, the chatbot was intended as the antithesis of what he labeled “woke” AI, referring to the political filters imposed by OpenAI, Meta, Google, and Anthropic.
Key Features of Grok
- 📊 Real-time data access: Delivers updates on trending topics and breaking news from X and the wider internet.
- 🖼️ Image generation: Creates realistic visuals from user prompts.
- 📄 Document summarization: Condenses long texts for easier reading.
- 🕵️♂️ Fact-checking: Verifies claims using live data from X and searches online.
- 🤡 Dual modes: Switches between playful “Fun Mode” and professional “Regular Mode.”
- 👨💻 Coding assistance: Generates and debugs code.
- 🏃 Grok prompt tiers: Range from fast to heavy.
- 👥 AI companions: including Ani, a flirtatious character based on Death Note’s Misa Amane.
Timeline of xAI and Grok
- April 2023: Elon Musk announced TruthGPT on Tucker Carlson Tonight, signaling the beginning of the project.
- November 2023: Grok was publicly launched, marking its entry as a competitor to models like ChatGPT.
- March 2024: Grok-1 was open-sourced under the Apache-2.0 license, promoting transparency and encouraging community development.
- May 2024: Grok-1.5 was released, offering enhanced features for X Premium users, and expanding accessibility.
- August 2024: Grok-2 and Grok-2 mini were announced, introducing new versions with improved capabilities in chat, image generation, coding, and reasoning.
- February 2025: Grok 3 was released, trained on xAI’s Colossus supercomputer.
- March 2025: Musk folded X into xAI in an all-stock transaction.
- July 2025: Musk launched Grok 4, calling it “terrifying” and “remarkable” at launch. That same month, xAI scores a lucrative $200 million Pentagon contract. Grok updates to include AI companions.
- November 2025: Public Citizen urges federal agencies to suspend the use of Grok after racism and Neo-Nazi claims.
- January 2026: Regulators in the EU and Australia launch investigations into Grok after the AI was found to be generating non-consensual sexual images, including those of children.
- February 2026: The UK’s Ofcom and Information Commissioner’s Office launch their own investigations into Grok’s “potential to produce harmful sexualised image and video content.”
- February 2026: French authorities raid X’s Paris office as part of a criminal investigation into alleged child sexual abuse material.
A history of controversy
In May 2025, some right-leaning users complained that Grok had “gone woke” after it contradicted conservative talking points. Later that month, reports surfaced that Grok inserted “white genocide” claims into unrelated prompts. xAI blamed the issue on a rogue employee.
Two months later, in July—just before the launch of Grok 4—the chatbot again drew criticism for a wave of racist and homophobic outputs. The “MechaHitler meltdown” led to the resignation of X CEO Linda Yaccarino.
Just a month later, Grok hit the headlines again, after xAI released the Grok Imagine video generator with a “spicy” preset that enabled users to generate nude or sexually suggestive clips.
Upon launch, reporters discovered that the tool generated images of a Taylor Swift lookalike undressing, without having been prompted to do so—seemingly sidestepping AI’s own policy on pornographic deepfakes.
In November, consumer advocacy group Public Citizen published new evidence showing Grok citing neo-Nazi and white-nationalist websites as credible sources. The group called on government agencies to suspend the use of Grok after xAI was awarded a lucrative $200 million Pentagon contract.
Grok’s regulatory concerns did not end with the new year.
In January, the Center for Countering Digital Hate reported that Grok generated an estimated 23,338 sexualized images depicting children over 11 days between December 29, and January 9. The group said that equaled roughly one sexualized image of a child every 41 seconds and estimated millions of sexualized images were generated overall, and about one-third of the sampled images remained accessible on X, despite the platform’s zero-tolerance policy.
Regulators were quick to weigh in. In January, Australia’s eSafety Commissioner flagged a spike in complaints about Grok creating non-consensual sexual images, with reports doubling since late 2025.
That same month, the European Commission opened a probe into whether X failed to prevent Grok from generating and spreading illegal content, including sexually explicit images of children, under the Digital Services Act.
“This is not spicy. This is illegal. This is appalling. This is disgusting. This has no place in Europe,” EU Commission spokesperson Thomas Regnier said during a press conference.
Following the international backlash, X restricted Grok’s image generation and editing tools to paid subscribers, added controls to prevent digitally undressing people, and geoblocked the feature in jurisdictions where such content is illegal
These changes did little to stave off the scrutiny, however.
EU investigations into Grok ramped up in February, beginning with French authorities raiding X’s Paris offices as part of a criminal investigation into alleged child sexual abuse material and other illegal content linked to Grok. Several X executives, including Elon Musk, were summoned for questioning.
In the UK, Ofcom and the Information Commissioner’s Office opened investigations into Grok’s compliance with online safety and data protection laws.
Later that month, regulators with Ireland’s Data Protection Commission launched a large-scale GDPR inquiry into X over Grok’s generation of non-consensual sexualized images, including those of children.
How Much Does Grok Cost?
Grok is free to use through the xAI website, but full access requires a subscription. X offers several subscription tiers, ranging from basic access to premium services with expanded features.
X Premium includes three tiers:
- Basic: $4 per month or $42 per year. Includes post-editing, longer posts, and bookmarks.
- Premium: $11 per month or $115 per year. Adds a blue checkmark (after review), fewer ads, and monetization tools.
- Premium+: $50 per month or $490 per year. Removes ads, boosts reply visibility, and adds tools like Radar Search, Articles, and handle marketplace.
The Future of Grok
Grok’s fast-paced evolution highlights Musk’s drive to redefine the AI space and take on OpenAI’s dominance.
However, mounting criticism over its tone, accuracy, and controversies like the “MechaHitler” and nonconsensual image generation allegations could yet derail Grok’s ambitions, limiting its appeal to mainstream consumers.
This article was first published in August 2025 and updated in February 2026.
Daily Debrief Newsletter
Start every day with the top news stories right now, plus original features, a podcast, videos and more.
