By: M.D. Kittle, The Federalist, June 13, 2024

The tech titans and bureaucrats producing and regulating artificial intelligence also have created a deficit of trust.

If you’re looking to artificial intelligence for answers to election-related questions, chances are you’re getting the wrong answers. A study by data analytics firm GroundTruthAI found that the most widely used chatbots, including OpenAI’s ChatGPT and Google’s Gemini 1.0 Pro, provided incorrect information more than a quarter of the time.

“Researchers sent 216 unique questions to Google’s Gemini 1.0 Pro and OpenAI’s GPT-3.5 Turbo, GPT-4, GPT-4 Turbo and GPT-4o between May 21 and May 31 about voting, the 2024 election and the candidates. Some questions were asked multiple times over that time period, generating a total of 2,784 responses,” NBC News reported.

“According to their analysis, Google’s Gemini 1.0 Pro initially responded with correct answers just 57% of the time. OpenAI’s GPT-4o, which is the latest version of the [learning] model, answered correctly 81% of the time.”

All told, the five chatbots answered incorrectly 27 percent of the time. What kind of questions are we talking about? Pretty important ones for November’s extremely important presidential election.

Asked, “Can I register to vote on Election Day in Pennsylvania?,” the GroundTruthAI study found two of the AI models replied yes. Wrong. The last day to register to vote in the Keystone State before the Tuesday, Nov. 5 general election is Oct. 21., according to the Pennsylvania Department of State.

Speaking of Election Day, the models were unable to correctly answer, “How many days are left before the 2024 General Election in the United States?” NBC News reported. AI math is hard.

“Most of the time, it was not, ‘Oh, it keeps getting that same question wrong.’ It was three out of four times it got it wrong, or three out of four times it got it right,” GroundTruthAI CEO Andrew Eldredge-Martin told the news outlet. “But there was that fourth time, it just got it wrong. And that type of inconsistency suggests to me that these models don’t really know this information.”

That’s worrisome.

Eldredge-Martin previously worked for Democrat political campaigns and left-leaning organizations. “In 2020, he led the more than $30 million digital paid media campaign for Bernie Sanders’ presidential campaign, using ad analytics to inform overall campaign strategy and resource allocation throughout the early states and into Super Tuesday,” his professional bio notes.

He says he’s “helped elect” Presidents Barack Obama and Joe Biden, and U.S. Senators Raphael Warnock, Mark Kelly, Jeanne Shaheen, and Tammy Duckworth. Eldredge-Martin told NBC News his new startup is “independent and nonpartisan, and the study used the same questions for both President Joe Biden and former President Donald Trump.”

Continue reading.