Oregon uses AI to flag election 'misinformation,' raising fresh concerns about censorship

The Oregon secretary of state’s office used artificial intelligence in the 2022 election that flagged election fraud concerns as mis-, dis-, and malinformation (MDM) and may use it again this year, similar to how Arizona monitored online election information. The effort is already facing a lawsuit.

Oregon flagged election information online during the 2022 midterms, with an AI targeting concerns regarding mail-in ballots. Arizona also worked with social media companies to take down social media posts that allegedly contained election misinformation. 

During the 2022 election cycle, the Oregon secretary of state at the time hired a U.K. artificial intelligence company for a pilot project “to help provide a suite of products to identify and disarm harmful information online as it relates to elections (mis-, dis-, and mal-information, or ‘MDM’),” according to a bid solicitation. 

Oregon State Rep. Ed Diehl (R) provided Just the News with documents he received from the secretary of state’s office via a public records request. 

The company, Logically AI, sent reports to the Oregon secretary of state regarding “MDM narratives” found online, including during the 2022 election. 

Some MDM narratives in 2022 that Logically AI flagged, according to its own report, were:

• “Mail-in ballots are used to commit voter and election fraud in Oregon;"
• “Future elections can’t be trusted due to the fraudulent 2020 elections;"
• And “voter and election fraud has been happening in Oregon for decades,”

There were also “high risk potential MDM narratives” flagged by Logically AI in August 2022, such as “Oregon should eliminate mail-in voting and hold in-person voting that requires voter ID” and “people should not trust mail-in ballots because signature verification is ineffective," the report said.

In 2005, the Commission on Federal Election Reform – a bipartisan panel that included ex-President Jimmy Carter – warned about issues with expanding mail-in voting and mentioned Oregon’s election process, while advocating for voter ID to prevent fraud. 

Last year, the secretary of state’s office posted another bid seeking a “solution” to combat MDM. 

State legislators sent the secretary of state a letter in November, calling on the secretary “to immediately halt implementation of the ‘Misinformation, Disinformation, and Mal-information (MDM) Analysis Platform Services’ Contract” because “the purpose of this system is to continually monitor and actively manipulate the free speech of Oregonians.” 

The secretary responded, saying that a contract hadn’t yet been signed with Logically AI after the pilot project, but that they “are currently in an RFP process for a system we’d like to bring online.” 

“The purpose of the proposed system will be to review publicly available information on websites, social media, and blogs, and notify us of threats and misinformation in a manner similar to a Google news alert,” the secretary added. “We will do two things with these notifications: 1) when there is a threat to life or infrastructure, we will notify the relevant law enforcement agencies, and 2) when there is false information about our elections, we will use our communications channels to share accurate information with voters. 

“My office has no authority to remove information from the internet, so there is no first amendment issue here. We do, however, have a very real need to protect the people and infrastructure that make our democracy work,” the letter continued. 

State legislators filed a lawsuit last month against state executives, requesting a preliminary injunction that would halt “any performance of work pursuant to" the anti-MDM contract. The case is currently pending. 

Earlier this month, the Oregon state House Republican Office sent questions to the secretary of state’s office related to Logically AI and monitoring information on social media platforms. 

According to answers sent from the secretary of state’s office, the “office reported instances of verifiably false information about election operations during the 2022 election. We ceased that practice after November 2022 because our reports were ignored by social media companies.

"We are not aware of a single instance in which our reports resulted in action by a social media company. Multiple administrations, the current SOS included, have reported threats to life and infrastructure to social media companies and as well as [sic] law enforcement," it added.

Diehl told Just the News on Friday about his concerns regarding the use of Logically AI by the secretary of state’s office. 

"When you start suppressing speech and do these actions, it feeds conspiracy theories, which is bad for the country," he said. 

The Oregon secretary of state's office and Logically AI didn't respond to requests for comment.

Oregon isn’t the only state to engage in monitoring election information on social media platforms. 

An email surfaced following the 2022 midterm elections that showed then-Arizona Secretary of State Katie Hobbs' office flagging a Twitter (now X) account on Jan. 7, 2021 for review. The message emerged during discovery in a First Amendment lawsuit filed in May by Missouri Attorney General Eric Schmitt and Louisiana Attorney General Jeff Landry against President Joe Biden alleging collusion between the administration and Big Tech regarding online censorship. 

Under the subject line "Election Related Misinformation," Hobbs' communications director cited two tweets from an account that were of "specific concern to the Secretary of State." 

In explaining the reason for the state intervention to seek suppression of the offending speech, the communications director said: 

"These messages falsely assert that the Voter Registration System is owned and therefore operated by foreign actors. ... This is an attempt to further undermine confidence in the election institution in Arizona."  

The email was sent to the nonprofit Center for Internet Security, which forwarded it to Twitter. "Both Tweets have been removed from the service" Twitter replied in an email copied to the Department of Homeland Security's Cybersecurity and Infrastructure Security Agency (CISA). 

After the information regarding Hobbs was revealed, a CISA report was posted on Twitter by Trump campaign lawyer Christina Bobb regarding Maricopa County Recorder Stephen Richer's March appearance before the DHS agency's Protecting Critical Infrastructure from Misinformation and Disinformation Subcommittee to brief the members on combating election misinformation, disinformation, and malinformation. 

According to the documents, Richer "suggested that CISA hold bootcamps for media representatives such as FOX News or CNN to enhance media's understanding of how elections are administered." He also said that the "malinformation" his office has encountered included "Abuse of Arizona's permissive public records process," as they received more than 350 public records requests last year regarding the 2020 election. 

Richer told the subcommittee that as election misinformation, disinformation, and malinformation continues to spread, "the pressure on his staff will continue to build and it will become difficult to perform statutory responsibilities needed to establish safe, credible, and fair elections."