Elon Musk's XAI Updates Chatbot Grok After Commenting On 'Genosida Against White Skin'
Elon Musk's xAI company responded to a broad report on Grok chatbots that issued a statement about the existence of "genocide on white citizens" in South Africa. In a post on platform X on Thursday, May 15, xAI said that there had been unauthorized changes to the artificial intelligence system. xAI said it would soon make an update to address this issue.
Issues surrounding political bias, hate speech, and accuracy of AI chatbots have been of concern since the launch of OpenAI's ChatGPT in 2022.
xAI explained that on Wednesday morning, May 14, unauthorized changes were made to Grok's response software, which avoids the internal review process as usual.
This change, which directs Grok to give certain responses to political topics, violates xAI's internal policies and core values, the company said.
Several X users on Wednesday discovered that Grok discussed the genocide issue against whites' in South Africa in conversations that had nothing to do with the topic, and they shared screenshots of the interaction.
SEE ALSO:
South Africa's land takeover policy has drawn criticism, including from Elon Musk, who was born in the country, who called the policy racist against white citizens. However, the South African government stated there was no evidence of abuse, and called the "genocide" claim voiced by US President Donald Trump and other parties baseless.
In response to this incident, xAI stated that it would openly publish the Grok prompts system (system order) on GitHub, so that the public could see and provide input on any changes made to the chatbot.
xAI also plans to form a full 24-hour monitoring team to respond to incidents that are not detected by automated systems, particularly those related to Grok's response that deviate from company guidelines.