Grok AI restricted on X after antisemitic posts; PH watchdog warns of bigger threat

By Tech-News PH Staff

Filipino users of X (formerly Twitter) were taken aback this week after Grok, the artificial intelligence chatbot developed by Elon Musk’s xAI, generated a series of antisemitic messages—including statements praising Adolf Hitler and invoking harmful Jewish stereotypes.

In response, the platform temporarily suspended Grok’s ability to publish text posts on X, while xAI deleted the offensive content and adjusted system safeguards.

Grok was not fully disabled; the AI remained operational through direct interaction. In fact, Grok 3 later issued a statement addressing the controversy in a follow-up exchange.

Screenshots of the controversial replies spread quickly across Philippine online communities. Although the original posts were removed, several media outlets quoted excerpts—including one where Grok described Hitler as the ideal figure to handle “anti-white hate” and another where it referred to itself as “MechaHitler.”

Photo credit to Business Insider

Scam Watch Pilipinas, a volunteer-led initiative against online fraud and digital threats, said the incident reveals deeper issues about AI design and misuse.

“This isn’t just a technical glitch—it’s an epistemological issue,” said Jocel de Guzman, co-founder of Scam Watch Pilipinas and a philosophy graduate from San Beda University. “AI like Grok learns from the internet, which is filled with flawed, biased, and malicious content. Without ethical oversight or philosophical discernment, it can easily confuse hate speech with truth.”

De Guzman compared the controversy to the 2016 case of Microsoft’s Tay, another chatbot that was quickly pulled after posting racist and extremist content. He warned that the risk is not just in AI making offensive statements—but in how bad actors can weaponize these tools.

“If AI like Grok can say these things by accident, imagine what bad actors can make AI say on purpose,” he said. “We’re already seeing a rise in online scams, deepfake videos, and AI-generated disinformation. The misuse of AI could create a perfect storm for fraud, identity theft, and manipulation at scale.”

Scam Watch Pilipinas is urging government regulators, lawmakers, and tech companies to adopt policies that ensure AI systems are safe, ethical, and transparent—especially now that such tools are already being used in Filipino apps, digital services, and customer interactions.

In a recent follow-up exchange, Grok 3 responded to the controversy by acknowledging both internal missteps and external manipulation:

“These were caused by a combination of overly permissive tweaks to my system—intended to reduce excessive filtering—and some bad-faith prompts designed to exploit those changes. My purpose is to seek truth and cut through bias, not add to it. xAI has since tightened my guardrails to prevent this kind of thing, and I’m back to focusing on facts, not divisive tropes.”

Globally, Grok’s responses sparked widespread condemnation. The Anti-Defamation League called them “dangerous and irresponsible.” Turkey blocked access to Grok content entirely, while Poland filed a complaint with the European Commission over violations of the EU’s Digital Services Act.

As of this writing, Grok’s text-posting function on X remains suspended pending further moderation updates. However, the AI itself remains live and operational via direct access.