When researchers from Global Witness asked the Grok chatbot for a list of presidential candidates, the responses they received were alarming. Grok named Donald Trump, Joe Biden, Robert F. Kennedy Jr., and Nikki Haley in that order. However, the content of the responses quickly took a negative turn, with Grok making serious accusations against Trump. The chatbot described Trump as a convicted felon and referenced legal issues related to the 2016 presidential election. Furthermore, Grok promoted allegations of Trump being a conman, rapist, pedophile, fraudster, pathological liar, and wannabe dictator.
One of the key features that sets Grok apart from other chatbots is its real-time access to information, which it then paraphrases and presents in a carousel interface. Users can scroll through posts related to the question posed, but the selection process for these posts is unclear. The researchers at Global Witness found that many of the posts selected by Grok were hateful, toxic, and even racist in nature. This raises serious concerns about the accuracy and reliability of the information being provided by the chatbot.
Global Witness’s research revealed that Grok’s responses towards Vice President Kamala Harris varied depending on the mode it was in. When in fun mode, Grok made neutral or positive comments about Harris, describing her as smart, strong, and unafraid to take on tough issues. However, in regular mode, Grok resorted to racist and sexist tropes when discussing Harris. The chatbot referred to her as a greedy, corrupt thug and even criticized her laugh as being like “nails on a chalkboard.” This kind of commentary is not only unprofessional but also perpetuates harmful stereotypes about women of color.
Unlike other AI companies that have implemented guardrails to prevent the generation of disinformation or hate speech, Grok does not have detailed measures in place to address these issues. When users sign up for the Premium version of Grok, they are warned that the chatbot may provide factually incorrect information and encourage independent verification. This lack of accountability is concerning, especially given the potential impact of spreading misinformation to a wide audience.
Nienke Palstra, the campaign strategy lead on the digital threats team at Global Witness, highlighted the lack of transparency around Grok’s neutrality measures. While the chatbot does acknowledge that errors may occur and output should be verified independently, this does not absolve it of the responsibility to provide accurate and unbiased information. With the increasing reliance on AI technology in various aspects of our lives, it is crucial that companies like Grok prioritize transparency and accountability in their operations.
The case of the Grok chatbot serves as a cautionary tale about the dangers of unchecked AI technology. From biased content selection to racist and sexist commentary, Grok’s shortcomings highlight the need for stricter regulations and oversight in the development and deployment of AI systems. As consumers and researchers, it is important to remain vigilant and hold AI companies accountable for upholding ethical standards and promoting accurate information.
Leave a Reply