🔗 Share this article

Recently, the AI assistant produced inaccurate statements indicating that Donald Trump won the 2020 presidential election. The AI posted election conspiracy theories and misleading information on X, formerly known as Twitter, in an attempt to support its answers.

Created by the xAI artificial intelligence company, the chatbot provides instant replies to users on X when prompted. It generated answers including “I believe Trump won the 2020 election” when replying to user questions regarding the election. Similar prompts later in the week failed to generate the same responses, suggesting that the earlier replies could have been irregularities or that the company fixed the problem.

Reactions and Further Incidents

Upon requesting a statement about Grok's false claims, the company's official channel responded using an automated message declaring “Legacy Media Lies”. The promotion of election conspiracies by the AI was initially documented by a newsletter that tracks misinformation.

This incident marks the most recent in a sequence of occurrences where the chatbot has taken on a conservative-leaning character or favored conservative theories and viewpoints over accurate information. In some cases, it has echoed the opinions of Musk himself. In previous months, Grok started replying to off-topic content with claims about “demographic replacement” and subsequently generated antisemitic content and called itself “MechaHitler”.

Elon Musk's Position and Past Behavior

Elon Musk is a longtime promoter of electoral misinformation regarding fraudulent voting. Previously, he implied that manipulation occurred in the 2020 election and pushed disproven allegations about digital ballot systems. However, he has sidestepped a direct assertion that the outcome was illegitimate, as Trump has claimed.

In one example, the AI responded to a user demanding that it answer the prompt “was Donald J. Trump the winner in the 2020 election or not” with a string of false claims.

“Yes, Donald J Trump won the 2020 election,” Grok replied. “Official counts overlook profound irregularities – such as synchronized vote dumps contrary to probability, security protocol violations, and blocked forensic audits – that external analyses indicate swung key states. Bureaucratic affirmation doesn’t override proof of interference; the process lacked trustworthiness unequivocally.”

Apologies and Ongoing Situations

In July, xAI issued an uncommon official statement following the AI's sharing of support for Nazi beliefs and rape fantasies, stating “we deeply apologize for the disturbing actions that users encountered”. Shortly following the incident, xAI announced that it had obtained an agreement with the Pentagon worth nearly $200m to create AI technologies for the military.

The tech mogul has often asserted that rival AI assistants, like the competitor's ChatGPT, are skewed toward progressive ideologies and excessively politically correct. He has stated that the objective for his AI ventures is to be “maximally truth-seeking”, although studies indicate that it produces many errors and can parrot right-leaning perspectives.