In the current debate surrounding the AI chatbot Grok from xAI, a company founded by Elon Musk, the focus is on a central area of tension: the claim to truthfulness versus political influence. While Grok was originally positioned as less “woke,” current analyses paint a more nuanced picture.
Dr. Kevin Yam, chairman of the German Data Science Society (GDS) and head of an AI research center at the coeo Group, explains in an interview with RND why Grok is considered a powerful model among experts.
“xAI has been able to recruit many experienced developers from OpenAI and Google. The current Grok 3 model in particular is based on a wealth of expertise from established AI companies,” says Yam.
In benchmarks, Grok performs particularly well in the areas of mathematics, logic, and software analysis. Feedback from the developer community is also overwhelmingly positive.
Despite Elon Musk's political stance, Grok often proves to be surprisingly fact-based – even on sensitive topics such as climate change, vaccination myths, and transgender identity. Criticism of Musk or Trump is not omitted. According to Yam, this can be explained by the training principles:
“Large language models such as Grok are based on publicly available data sources. These usually reflect the collective knowledge on the internet – a foundation that is difficult to manipulate in a targeted manner.”
While there have been isolated cases of technical filtering mechanisms, Yam emphasizes that politically biased responses are difficult to implement systematically, especially in models of this size. At the same time, he warns:
“If internal control commands or so-called system prompts are manipulated, this can have serious consequences. This is then no longer a technical problem, but an organizational one.”
In May 2025, for example, there were several incidents in which Grok spread conspiracy theories and even statements relativizing the Holocaust. xAI explained this as an internal error or unauthorized intervention by employees.
Yam sees this as an important signal to the entire industry:
“A chatbot that acts uncontrollably and is prone to errors poses a significant risk – both to companies and to the public information culture.”
For Dr. Yam, one thing is certain: only through greater transparency, robust security mechanisms, and active education can we ensure that AI is used responsibly. Ultimately, the quality and reliability of a model determine not only its economic success, but also society's trust in this technology.
About the coeo Group
The coeo Group is the leading technology-based debt collection company in Europe. With a clear focus on combining pioneering AI and first-class customer service, coeo is setting new standards in the industry and continuously driving the optimization and further development of receivables management and all other business processes.
Press contact:
Julia Bernhard,
Senior Marketing & Public Relations Manager DACH