The social media platform is modifying Grok following a criticism that it produced false details about voting deadlines in US states.
Social media platform X has made a change to its synthetic intelligence chatbot after 5 US secretaries of state warned it was spreading election misinformation.
High election officers in Michigan, Minnesota, New Mexico, Pennsylvania and Washington despatched a letter this month to Elon Musk complaining that the platform’s synthetic intelligence chatbot, Grok, produced false details about state voting deadlines shortly after President Joe Biden dropped out of the 2024 presidential race.
The secretaries of state requested that the chatbot direct customers asking election-related inquiries to CanIvote.org, an election info web site run by the Nationwide Affiliation of Secretaries of State.
Earlier than itemizing solutions to election-related questions, the chatbot now says: “For correct, up-to-date details about the 2024 U.S. elections, please go to Vote.gov.”
Each web sites are “trusted sources that may join voters with their native election officers,” the 5 state secretaries mentioned in a shared assertion.
“We admire X’s motion to enhance its platform and hope that they are going to proceed to make enhancements that can guarantee their customers have entry to correct info from trusted sources on this crucial election 12 months,” they mentioned.
Grok is simply accessible to subscribers of premium variations of X. However the secretaries of state who signed the letter mentioned that election misinformation from Grok has been shared throughout a number of social media platforms and reached tens of millions of individuals.
Grok continued to repeat the false info for 10 days earlier than it was corrected, the secretaries mentioned.
The platform didn’t reply to a request for remark.
The change selling a hyperlink to an official voting web site doesn’t seem to handle Grok’s potential to create deceptive AI-generated photos associated to elections. Individuals have been utilizing the instrument to flood the platform with pretend photos of candidates, together with U.S. Vice President Kamala Harris and former President Donald Trump.
Grok debuted final 12 months for X Premium and Premium Plus subscribers, and Musk touted it as a “rogue” AI chatbot that can reply “spicy questions which are rejected by most different AI techniques.”
Social media platforms have confronted rising scrutiny for his or her function in spreading misinformation, together with about elections.
The letter additionally warned that inaccuracies may be anticipated in AI merchandise, particularly chatbots like Grok, which depend on massive language fashions.
Since Musk purchased Twitter in 2022 and rebranded it as X, watchdog teams have raised considerations a couple of rise in hate speech and misinformation being amplified on the platform, in addition to cuts to workers who moderated content material.
