On Monday, a group of top election officials from five states wrote a letter urging billionaire Elon Musk to put an end to the spread of election misinformation on X, which was created as an AI chatbot by him.
In a joint letter, secretaries of state from Minnesota, Pennsylvania, Michigan, Washington state, and New Mexico demanded that Musk take immediate action to implement changes to X’s AI search assistant, Grok. The purpose of the changes is to ensure voters have access to accurate information during this critical election year.
In November of last year, Musk introduced Grok on X, labeling it as a chatbot that provides unfiltered responses which differ from those of large language models such as ChatGPT. He mockingly referred to companies like OpenAI and Google as “woke” for their use of guardrails to assist the tools in approaching sensitive and controversial topics more cautiously.
During the unveiling of Grok, xAI – Musk’s AI firm, made it clear that the tool should only be used by those who appreciate humor.
The secretaries of state have expressed their apprehension about Grok’s involvement in the dissemination of false information regarding the 2024 presidential election. They cited an incident from last month when the chatbot generated inaccurate details about ballot deadlines, which came just a few hours after President Joe Biden announced that he wouldn’t run for re-election.
In a recent post by Grok, it was falsely stated that Kamala Harris, the presumptive Democratic nominee and Vice President, had failed to meet the ballot deadline for the upcoming November election in several states including Alabama, Indiana, Michigan, Minnesota, New Mexico, Ohio, Pennsylvania, Texas, and Washington. These states are crucial battleground areas that could play a significant role in determining the outcome of the election.
The letter strongly refuted Grok’s claim by stating, “This is false.” It further added that the opposite is true in all nine states. The ballots are not closed yet, and the upcoming deadlines would allow changes to the candidates listed on the ballot for the offices of President and Vice President of the United States.
Currently, Grok is exclusively accessible to X Premium and Premium+ subscribers. The bot comes with a disclaimer, urging users to verify the information it produces. Despite this, the misinformation about ballot deadlines from Grok found its way into more publicly accessible posts, reaching millions, as reported by the secretaries of state. It took 10 days for the chatbot’s misinformation to be corrected.
The letter addressed to X highlights the crucial responsibility of providing accurate information about voting to the tens of millions of voters in the US during this major election year. According to The Washington Post, the letter emphasizes the need for X to ensure that voters using their platform have access to genuine guidance on their constitutional right to vote. It is imperative that X upholds its duty in providing reliable information to its users.
The secretaries of state have recommended X to adopt a new policy that would guide Grok users to “CanIVote.org” when seeking information about U.S. elections. According to the officials, this website is a reliable and unbiased resource created by election administrators from both Democratic and Republican parties. In collaboration with the secretaries of state, OpenAI has been working to improve the accuracy of election information, and ChatGPT has already been integrated to redirect users to this website whenever they inquire about the elections.
Despite HuffPost’s request for comment, X’s spokesperson did not respond. Instead, an automatic response was received stating that they were currently busy and to check back later.
According to Minnesota Secretary of State Steve Simon, the company’s response to Grok’s misinformation was delayed and essentially amounted to a shoulder shrug. He was the driving force behind the letter.
Simon emphasized the significance of social media companies, particularly those with worldwide influence, taking responsibility for their mistakes. He cited the example of the Grok AI chatbot, which made errors due to misunderstanding the guidelines. According to Simon, speaking up now could potentially lower the likelihood of social media companies ignoring or postponing corrections of their own errors leading up to the November election. Simon stated this to the Post.
