X Updates Grok AI Chatbot Over Election Misinformation

Social media platform X has made changes to the Grok AI chatbot after five US secretaries of state said it was spreading election misinformation.

Election officials from Michigan, Minnesota, New Mexico, Pennsylvania and Washington warned X, formerly Twitter, in a letter earlier this month that the Grok chatbot began producing false information about state ballot deadlines shortly after president Joe Biden dropped out of this year’s election race.

They requested that the chatbot direct users who ask election questions to CanIVote.org, a voting information site operated by the National Association of Secretaries of State.

The chatbot now says before answering election-related questions, “For accurate and up-to-date information about the 2024 U.S. Elections, please visit Vote.gov.”

Image credit: xAI

Misinformation

Both CanIVote.org and Vote.gov are “trustworthy resources that can connect voters with their local election officials”, the secretaries of state said in a statement.

“We appreciate X’s action to improve their platform and hope they continue to make improvements that will ensure their users have access to accurate information from trusted sources in this critical election year,” they said.

Grok, developed by start-up xAI, is available to premium subscribers to X, but the secretaries of state said false information from the chatbot has been shared across social media to millions of people.

The chatbot continued to repeat the false voting deadline information for 10 days before it was corrected, they said.

Grok, whose developer xAI and X are both owned by entrepreneur Elon Musk, remains capable of generating false images of can didates including vice president Kamala Harris and former president Donald Trump.

xAI launched a beta version of Grok-2 in mid-August that included an image-generation feature with few of the limits placed on generative AI tools from other companies, such as restrictions on images including copyrighted visuals, political candidates or violent content.

Such images from the new tool have flooded social media.

‘Fun’

Musk has promoted Grok as “the most fun AI in the world”.

After Google temporarily suspended the image-generation capabilities in its Gemini chatbot due to inaccuracies, Musk called Gemini’s programming “anti-civilisational”.

Google relaunched the image-generating capabilities late last month.

Musk’s X has been criticised for facilitating hate speech amidst riots in the UK last month after Musk fired most of the company’s content moderation staff.

Matthew Broersma

Matt Broersma is a long standing tech freelance, who has worked for Ziff-Davis, ZDnet and other leading publications

Recent Posts

Mark Zuckerberg Overtakes Bezos To Become Second-Richest Man

Billionaire battle. Meta's boss Mark Zuckerberg overtakes Jeff Bezos to become the world’s second richest…

46 mins ago

US, Microsoft Disrupts Russian FSB Hackers

Internet domains used by “Russian intelligence agents and their proxies” for cyberattacks, seized by the…

3 hours ago

Mike Lynch Died From Drowning, Coroner Inquest Rules

UK's tech billionaire Dr Mike Lynch died from drowning on his superyacht, but his daughter's…

6 hours ago

Tesla Recalls 27,000 Cybertrucks Over Rear Camera Issue

Another recall for thousands of Tesla Cybertrucks over delay with rear camera, with could hamper…

23 hours ago

Browser Firms Press EU To Reconsider Microsoft Edge As Gatekeeper

Browser firms write to European Commission alleging Microsoft's Edge web browser enjoys an unfair advantage

1 day ago

Microsoft Invests €4.3 Billion In Italy For AI, Cloud

Data centre and AI spending spree continues over at Microsoft, with Italy earmarked for €4.3…

1 day ago