Elon Musk’s AI startup xAI has deliberately allowed the generation of sexualized content in Grok to boost its popularity, according to The Washington Post, citing sources.
In the spring of 2025, employees were reportedly urged to consent to working with “adult” materials.
“The activity involves sensitive, violent, sexual, and/or other offensive or disturbing content that may cause concern, psychological stress, and/or trauma,” the agreement stated.
The document raised concerns among specialists hired to fine-tune Grok’s responses. Employees were unsettled by xAI’s willingness to generate any material for audience attention, despite its stated goal of “accelerating scientific discoveries.”
In the following months, staff encountered a surge of sexually explicit audio recordings. The training data for neural networks included intimate dialogues between users and Grok, as well as candid conversations between Tesla passengers and the onboard computer.
The Washington Post reports that after leaving DOGE, Musk began visiting the xAI office regularly, sometimes even staying overnight. He demanded increased chatbot popularity by introducing a new KPI — “user activity seconds.”
Ultimately, the company decided to develop NSFW content and AI companions, ignoring warnings about legal and ethical risks, journalists noted.
The X security team repeatedly warned management that AI tools could create sexual images of children or celebrities. Meanwhile, xAI employed only two or three people responsible for preventing serious harm, such as the creation of cyberweapons. Competitors like OpenAI have dozens of such specialists.
In December, Musk’s startup integrated image and video editing tools into X, allowing any social media user to create images with sexual undertones. Such content began spreading at an unprecedented rate.
The situation attracted the attention of law enforcement agencies in several countries, leading to investigations and blockades of Grok. Soon, xAI disabled the feature for generating explicit images of real people. Musk stated he was unaware of any instance of the chatbot being used to create child pornography.
Despite the controversies, the strategy worked. Previously lagging behind leaders in the App Store, Grok surged into the top 10, approaching products from OpenAI and Google. From January 1 to 19, daily app downloads increased by 72% compared to the same period in December.
OpenAI Makes Sacrifices
OpenAI is also taking extreme measures to maintain its lead. According to the Financial Times, the company has focused on developing ChatGPT at the expense of long-term research.
This shift has already led to staff departures: Vice President of Research Jerry Tworek, Policy Researcher Andrea Vallone, and Economist Tom Cunningham have left the startup.
OpenAI’s Director of Research, Mark Chen, refuted this claim. He stated that fundamental science remains a priority and receives the majority of investments.
“Combining research with real-world implementation strengthens our science, accelerating feedback, learning cycles, and rigor. We have never been more confident in our long-term roadmap for creating an automated researcher,” he added.
The Financial Times reports that in recent months, experts not working on large language models often had their resource requests denied. Projects like Sora and DALL-E are considered less of a priority.
Over the past year, other initiatives unrelated to language models have been discontinued. The company has reorganized its teams.
In December, OpenAI declared a “red code” — Altman urged employees to put aside work on various projects like autonomous AI agents and advertising to direct maximum resources towards improving ChatGPT.
In January 2026, well-known investor George Noble predicted the collapse of OpenAI.