{"id":94693,"date":"2025-05-16T04:12:05","date_gmt":"2025-05-16T04:12:05","guid":{"rendered":"https:\/\/neclink.com\/index.php\/2025\/05\/16\/xai-blames-groks-obsession-with-white-genocide-on-an-unauthorized-modification\/"},"modified":"2025-05-16T04:12:05","modified_gmt":"2025-05-16T04:12:05","slug":"xai-blames-groks-obsession-with-white-genocide-on-an-unauthorized-modification","status":"publish","type":"post","link":"https:\/\/neclink.com\/index.php\/2025\/05\/16\/xai-blames-groks-obsession-with-white-genocide-on-an-unauthorized-modification\/","title":{"rendered":"xAI blames Grok’s obsession with white genocide on an ‘unauthorized modification’"},"content":{"rendered":"


\n<\/p>\n

\n

xAI blamed an \u201cunauthorized modification\u201d for a bug in its AI-powered Grok chatbot that caused Grok to repeatedly refer to<\/a> \u201cwhite genocide in South Africa\u201d when invoked in certain contexts on X.<\/p>\n

On Wednesday, Grok began replying to dozens of posts on X with information about white genocide in South Africa, even in response to unrelated subjects. The strange replies stemmed from the X account for Grok, which responds to users with AI-generated posts whenever a person tags \u201c@grok.\u201d<\/p>\n

According to a post Thursday from xAI\u2019s official X account, a change was made Wednesday morning to the Grok bot\u2019s system prompt \u2014 the high-level instructions that guide the bot\u2019s behavior \u2014 that directed Grok to provide a \u201cspecific response\u201d on a \u201cpolitical topic.\u201d xAI says that the tweak \u201cviolated [its] internal policies and core values,\u201d and that the company has \u201cconducted a thorough investigation.\u201d <\/p>\n

\n
\n
\n

We want to update you on an incident that happened with our Grok response bot on X yesterday.<\/p>\n

What happened:
On May 14 at approximately 3:15 AM PST, an unauthorized modification was made to the Grok response bot’s prompt on X. This change, which directed Grok to provide a\u2026<\/p>\n

\u2014 xAI (@xai) May 16, 2025<\/a><\/p><\/blockquote>\n<\/div>\n<\/figure>\n

It\u2019s the second time xAI has publicly acknowledged an unauthorized change to Grok\u2019s code caused the AI to respond in controversial ways.<\/p>\n

In February, Grok briefly censored<\/a> unflattering mentions of Donald Trump and Elon Musk, the billionaire founder of xAI and owner of X. Igor Babuschkin, an xAI engineering lead,\u00a0said that Grok had been instructed by a rogue employee<\/a> to ignore sources that mentioned Musk or Trump spreading misinformation, and that xAI reverted the change as soon as users began pointing it out.<\/p>\n

xAI said on Thursday that it\u2019s going to make several changes to prevent similar incidents from occurring in the future. <\/p>\n

Beginning today, xAI will publish Grok\u2019s system prompts<\/a> on GitHub as well as a changelog. The company says it\u2019ll also \u201cput in place additional checks and measures\u201d to ensure that xAI employees can\u2019t modify the system prompt without review and establish a \u201c24\/7 monitoring team to respond to incidents with Grok\u2019s answers that are not caught by automated systems.\u201d<\/p>\n

Despite Musk\u2019s frequent warnings of the dangers of\u00a0AI<\/a>\u00a0gone<\/a>\u00a0unchecked<\/a>, xAI has a poor AI safety track record. A recent report<\/a>\u00a0found that Grok would undress photos of women when asked. The chatbot can also be\u00a0considerably more crass<\/a>\u00a0than AI like Google\u2019s Gemini and ChatGPT, cursing without much restraint to speak of.<\/p>\n

A study by SaferAI, a nonprofit aiming to improve the accountability of AI labs, found xAI ranks poorly on safety among its peers, owing to its\u00a0\u201cvery weak\u201d risk management practices<\/a>. Earlier this month, xAI missed a self-imposed deadline<\/a>\u00a0to publish a finalized AI safety framework.<\/p>\n<\/div>\n