Secretary of State Caught Using AI to Flag Accurate Vote-Fraud Stories as ‘Misinformation’


(TJ Martinell, The Center Square) An artificial intelligence company working on behalf of Washington Secretary of State Steve Hobbs’s office to track election and voter “misinformation” flagged factually accurate stories on noncitizens voting that were published by the Center Square.

While the reports from the AI company claimed the stories could undermine confidence in the state’s election process, the Center Square’s review of the stories were based on internal government documents and emails obtained through public-records requests.

Last year, the SOS signed a contract with U.K.-based AI company Logically to search online content on various social media platforms, including Twitter, for “harmful narratives” regarding Washington’s elections or voting.

The company was also tasked with generating reports to be sent to SOS for review highlighting “Emerging Narratives” related to Washington’s elections or voting.

According to SOS, “The criteria for these searches was specified by OSOS staff and is nonpartisan.”

In June and July, Logically generated reports on two separate stories published by the Center Square, which obtained the reports:

The first report claimed that the first story pushed the narrative that “no government agency in Washington State has the mandate or authority to verify citizenship of registered voters and some users claimed it represents intentional negligence by Washington authorities to enable voter fraud. Users claimed that this leads to foreign nationals regularly voting in Washington’s elections.”

The report warned that the story could “motivate individuals to call for stricter voter registration laws and more oversight of the process.”

The second report claimed that a story on a foreign national avoiding prosecution after voting 28 times meant “that Washington’s elections were not secure and should not be trusted, and they noted that the sheriff’s office declined to prosecute the foreign nationals.”

Though the report included screenshots of online users making such comments along with posts of the Center Square stories, the report also included social media posts simply linking to the story itself without additional commentary.

The report further warned that the Center Square story could “increase claims of voter fraud and motivate individuals to call on election officials to implement citizenship verification procedures for voter registration in Washington State.”

When The Center Square reached out to the SOS inquiring which state agency verifies citizenship of registered voters, Deputy Director of External Affairs Derrick Nunnally wrote that “voters are required to affirm eligibility when they register (on paper, in person, or online) and with their signature each time they cast a ballot.”

Nunnally claimed that the state imposed tough consequences on those who violated the law and that the rolls were regularly updated using national databases like the Electronic Registration Information Center.

“Non-citizens who attempt to participate in elections can be subject to prosecution, deportation, and/or potential loss of a path to citizenship,” he claimed. “Washington has 4.8 million active voter registrations, which are regularly verified using ERIC and other state and federal data sources.”

The SOS did not directly respond to questions regarding actions taken by the SOS to ensure non-citizens do not vote in elections and the coming primary.

Nunnally also stated that the Logically AI reports sent to the SOS were used “to formulate the office’s social media posts and other messaging” rather than to pressure companies like Twitter to shadow-ban the offending outlets.

“Neither Logically AI nor the Office of Secretary of State has used these reports to seek censorship, categorize specific social media users, or request action against any social media user,” he claimed.

Twitter, Facebook and other outlets long denied such censorship practices in the leadup to the 2020 election, although it was later revealed through internal communications exposed by the Twitter Files and elsewhere that the companies were colluding with deep-state intelligence agencies to do just that.

Whether Washington State’s office is doing it or not, the misuse of AI in such a way raises concerns about leftist censorship efforts on a larger scale. The U.S. State Department, for instance, was revealed in a recent lawsuit to be colluding with private activist groups such as the Global Disinformation Index and NewsGuard to establish a de-facto blacklist of mostly conservative outlets who promoted information that they deemed problematic.

When contacted by the Center Square about the SOS program, Aaron Terr, director of public advocacy at the Foundation for Individual Rights and Expression, wrote that they were “concerned about how government efforts to combat misinformation can give rise to censorship.”

The actions being taken by the Washington State officials suggested, at best, that they were going down a slippery slope that violated the spirit of the First Amendment.

“Government monitoring of publicly accessible social media posts for alleged misinformation doesn’t in itself violate the First Amendment, even if it gives off a creepy authoritarian vibe,” Terr wrote.

“But First Amendment issues could arise depending on what the government does with that information,” he added. “We’ve seen, for example, how White House officials have unlawfully coerced social media platforms to remove or downgrade content related to controversial topics like COVID-19 and election fraud.”

Terr suggested that additional transparency and accountability on the part of Hobbs’s office and others attempting to use this technology would help to reassure free-speech watchdogs and members of the public that they were acting in good faith.

If “government officials are only trying to understand prevailing narratives about an issue so they can respond with their own speech, that’s less concerning,” he said.

“But the government should always be fully transparent about its actions and careful not to do or say anything to suggest citizens will face repercussions for constitutionally protected speech, even if officials deem that speech false or misleading,” he added. “The First Amendment doesn’t permit the government to dictate truth and squash dissent.”

Headline USA’s Ben Sellers contributed to this report.