The Grok chat bot developed by xAI stands out with its restrictions fewer compared to other artificial intelligence companies. According to a new report, this strategy directly affects a large portion of the company’s user traffic.

According to the report published by The Information, more than half of the traffic on Grok consists of NSFW, that is, inappropriate content. Information based on two former xAI employees reveals that users extensively use the chatbot to produce erotic content and adult role-playing chats.

Unexpected Trends in Grok Usage

It is stated that users discovered that models designed for writing code were more cost-effective and used these models to produce pornographic content. Internal analysis shows that a significant portion of the requests coming to the coding model are requests for nude images or pornographic content.

SpaceX has informed investors that Grok’s more unruly features could pose a risk. The company announced that it allocated a budget of $ 530 million for possible legal processes.

Internal Challenges and Future Risks

While xAI engineers are trying to manage users’ sexual chat requests, they are also working hard to block illegal content such as child abuse. It is said that there is no technical quick solution to this situation.

It is stated that some employees working on Grok’s anime style avatar Ani were uncomfortable with these processes. In addition, scandals regarding the creation of sexually explicit images of real individuals and children have created significant unrest within the company.

Although the creation of unsolicited content with images of real people has been restricted by some regulations made directly on the X platform, tests have confirmed that paid subscribers can continue to produce such materials.

xAI’s engagements with government agencies and the military could pose new challenges in the future for this kind of content-focused AI model. It now remains unclear how relations with the federal government will be affected by these controversial events.

Do you think allowing artificial intelligence models to allow this type of content is a necessary strategy for the commercial success of companies or is it a big risk?