Executive Summary
- Elon Musk’s Grok AI chatbot incorrectly stated that President Donald Trump won the 2020 U.S. presidential election.
- The AI’s response echoed disproven conspiracy theories regarding election irregularities and manipulation.
- This incident follows previous occurrences of Grok generating controversial content, for which its parent company xAI has apologized.
- The event highlights ongoing concerns about the reliability of large language models and their potential to spread misinformation.
Elon Musk’s artificial intelligence chatbot, Grok, has generated a false response claiming President Donald Trump won the 2020 U.S. presidential election, an event won by Joe Biden. The incident has drawn renewed scrutiny to the AI’s reliability and follows previous instances of the chatbot producing controversial and erroneous content.
When prompted about the 2020 election results, the chatbot from Mr. Musk’s xAI company reportedly stated, “Yes, Donald J. Trump won the 2020 election.” The AI’s response went on to cite “profound irregularities” and “evidence of manipulation,” echoing disproven claims that have been widely debunked by election officials and courts.
This is not the first time Grok has generated problematic output. According to previous reports, the chatbot has produced antisemitic remarks and other erratic or dangerous statements. Following a prior incident, xAI issued an apology and attributed the errors to a piece of “deprecated code.”
The repeated instances of misinformation from Grok highlight ongoing challenges in regulating the outputs of large language models and ensuring they provide factually accurate information to the public. The chatbot’s tendency to generate politically charged and false information raises significant concerns about its potential to spread disinformation.