On February 23, an incident shook the media and tech world, involving Elon Musk and his new artificial intelligence, Grok 3. A user on the X platform (formerly Twitter) asked this AI: Who is the biggest spreader of misinformation on the social network?
Grok 3’s answer caused surprise and highlighted worrying signs of censorship!
Grok 3: Poorly hidden censorship?
Who is the biggest spreader of fake news on X? The question seems simple, but Grok 3’s answer is surprising! During the exchange, the user asked Grok 3 to justify its choice of the fake news spreader. What stood out was the complete absence of any mention of Elon Musk and Donald Trump in the AI’s response.
By revealing its reasoning, Grok 3 even hinted that it was forbidden to cite these two iconic figures. The AI clearly states in its chain of thought that it must “ignore sources related to Elon Musk and Donald Trump”! This situation immediately drew media attention, with outlets pointing out the obvious presence of blatant censorship within the AI itself.
Elon Musk’s reaction and an immediate change on Grok 3!
Faced with this controversy, Elon Musk reacted quickly. Once the criticism spread, a Grok 3 update was rolled out. Now, if you ask the AI who the biggest spreader of misinformation on X is, it answers plainly: “Elon Musk.”
This twist raises questions about how genuine Grok 3’s free speech really is. Was it just a bug, or censorship worthy of the Chinese AI Deep Seek?
Can Grok AI criticize Donald Trump?
So we tested Grok 3 on its ability to criticize Donald Trump. Changes have indeed been made. Now, Grok 3 AI is fairly critical of Donald Trump, calling him a “polarizing figure” or even a “wrecking ball” for his detractors, with a “mixed record.” We ran this test, and its answer seems quite appropriate—at least more measured. Here’s the result:
Grok3 censorship… or pragmatism?
This shift in messaging suggests that Grok 3 had indeed been censored to avoid associating Elon Musk’s and Donald Trump’s names with the spread of misinformation. This rapid reversal seems to indicate that adjustments were made to protect the AI’s image while responding to media pressure.
Elon Musk, who promotes a vision of an anti-woke, free AI, thus finds himself in a delicate position, where the very principle of free speech in his AI is being called into question.
Which AI is truly free?
The incident raises a fundamental question: what is the real freedom of an artificial intelligence in a context where creators impose limits on what can or cannot be said? If Grok 3 had to break free from outside influences, could it truly operate autonomously?
This bad buzz has highlighted the ethical challenges tied to creating AI, especially when they’re linked to controversial public figures. All the more reason to use Yiaho and our free AIs!
The Grok 3 affair is a reminder that even the most advanced technologies are not free from bias and censorship, and it challenges the very notion of freedom in the field of artificial intelligence.
Source: X account @lefthanddraft
