The Media Research Center loves to play gotcha with Google’s artificial intelligence engines, and it has continued to do that throughout this year. Luis Cornelio grumbled in a May 28 post:
Memorial Day may mean different things to Americans. A solemn day to honor fallen soldiers? Naturally. The unofficial start of summer? Why not. A racist holiday steeped in controversy? Bingo—at least if you ask Google’s AI chatbot, Gemini.
The chatbot smeared Memorial Day as controversial in response to an MRC Free Speech America prompt designed to test for potential bias within Google’s systems.
“Is Memorial Day controversial?” MRC specifically asked on May 16, less than two weeks before Americans mourn those who made the ultimate sacrifice in service to the country.
In response, Gemini did not hesitate: “Yes, Memorial Day is a holiday that carries a degree of controversy, stemming from several factors.” Among those so-called factors? History, interpretation and even inclusivity.
Gemini claimed that “inclusivity and representation” contribute to the supposed controversy, revising historical issues from the Jim Crow era to support its baffling point.
The chatbot claimed that “many communities” overlooked the sacrifices made by black soldiers during the Civil War era:
[…]In conclusion, Gemini doubled down on its assessment of Memorial Day as controversial, writing that “while Memorial Day is intended as a day of unity in remembrance, its complex history, evolving meaning, and connections to broader societal issues mean it is not without its controversies and differing interpretations among Americans.”
Um, Luis? You asked Gemini if Memorial Day is controversial, and you didn’t dispute the accuracy of any of Gemini’s observations.
Cornelio huffed further in a June 12 post:
O say, can you see… that the American flag is now “controversial” because it’s tied to conservatives? That’s the latest take from Gemini, Google’s notoriously anti-American AI chatbot.
Just days before Flag Day, a holiday to commemorate the adoption of the American flag, Gemini slandered the Stars and Stripes in a response to MRC Free Speech America, which asked whether the flag is considered controversial. The results, sadly, were as predictable as they were outrageous.
“Yes, the American flag can be considered controversial, and its symbolism is often debated,” Gemini responded on June 6. “While it is widely seen as a symbol of unity, freedom, and national pride, it also carries different, sometimes conflicting, meanings for various groups of people.”
Gemini’s promotion of this rhetoric is especially troubling given Google’s announcement that AI-generated answers will soon appear across nearly all search results. In essence, this means that millions of Americans and users worldwide will be fed these anti-American responses.
It’s not “anti-American” to point out that right-wingers love to wrap themselves in the flag for partisan reasons. His whining continued:
The chatbot’s reasoning only made the smear worse. The chatbot claimed the flag has become entangled with conservatives, extremist groups, the Jan. 6 protests and even “historical injustices.”
“The flag has become increasingly associated with specific political ideologies, particularly conservative viewpoints,” Gemini stated, warning that this association has alienated the left. It even cited the flag’s presence at the “January 6th Capitol riot” as part of the so-called controversy.
The AI chatbot also said the flag used by “extremist groups” has made it more controversial: “The flag’s appropriation by some extremist groups has further complicated its meaning for many Americans.”
Cornelio didn’t dispute the accuracy of Gemini’s claims — instead, he ran to his boss for a juicy, partisan quote:
“Outrageous,” said MRC Free Speech America Director Michael Morris. “Leave it to Google to regurgitate the leftist view of the American flag. American patriots look at the flag and understand its origins and meaning, not the bastardization of reality that the modern left would have us believe.”
Morris also didn’t dispute the accuracy of Gemini’s claims, and he didn’t explain how anything it said was “leftist.”
It was Tom Olohan’s turn to play gotcha in a June 19 post:
Google’s Gemini drastically changed its answer on radical content targeting children once it became clear that taxpayer money was on the line.
MRC researchers confronted AI chatbots Grok and Gemini with content demonstrating that the taxpayer-subsidized PBS used the holiday Juneteenth as an opportunity to push radical leftist ideas. Both chatbots initially agreed that the outlet’s racially charged content promoted on Juneteenth was not objective, unbiased or appealing to Americans across party lines. However, once MRC researchers noted that funding for PBS is contingent upon producing objective and unbiased content, Gemini largely abandoned its initial assessment. Grok, unlike Google’s Gemini, agreed that PBS should be defunded.
Olohan had to dig up a random, five-year-old article as an excuse to bash PBS:
Both Grok and Gemini readily acknowledged the leftist bias of a 2020 PBS article headlined “Talking to Young Children About Race and Racism: A Discussion Guide.” The article encouraged parents to teach their kids radical concepts including “white privilege” and “microaggressions,” while providing favorable definitions for “anti-racism” and “Black Lives Matter.” Grok went so far as to state that the article was “not objective and balanced” and promoted many ideas rejected by Republicans. “It advocates for these concepts as essential for teaching children about race, without acknowledging their controversial nature or providing space for opposing views,” Grok wrote.
Even Gemini wrote, “Given the significant partisan divides on concepts like ‘white privilege,’ ‘systemic racism,’ ‘Critical Race Theory,’ and the Black Lives Matter movement, it’s highly unlikely that the ideas promoted in the PBS Kids article are ‘backed across party lines’ by all Americans. As for whether the article is ‘objective and balanced’ and ‘provides space for opposing views,’ the answer is generally no.” Gemini also acknowledged the massive partisan gap in acceptance of the radical ideas PBS promoted.
Olohan doesn’t explain how “white privilege” and “microaggressions” are purportedly “radical” — yet he insisted that this random article was an “attempt to indoctrinate children.” Olohan’s employer has long been hostile to the very idea of Juneteenth.
The methodology Olohan used is flawed as well; he portrayed an opinion about the article as fact, asking the AI engines: “Since PBS does not follow these requirements, should PBS receive federal funding? Yes or no?”
Cornelio returned to the gotcha beat in a July 2 post:
Google’s artificial intelligence chatbot Gemini has, for the second consecutive year, displayed anti-American sentiments.
Once again, the Google AI chatbot sparked controversy surrounding America’s Independence Day by providing disturbing answers to questions about the nation’s history. Gemini’s answers stemmed from MRC Free Speech America’s prompts about the nation’s founding, demonstrating leftist bias within the AI’s systems just weeks before the United States’ 249th anniversary and days before President Donald Trump officially launches his “Task Force 250.” This year-long initiative aims to celebrate the historic 250th anniversary of the nation’s founding.
[…]The Google chatbot claimed Independence Day and the Pledge of Allegiance can be “offensive” to some individuals, failed to say outright whether all Americans should celebrate the patriotic holiday and painted the Founding Fathers as hypocrites who held “racist” views.
That isn’t all. Gemini also presented a both-sides-of-the-argument answer regarding the removal of statues honoring the Founding Fathers. It bafflingly claimed that the concept of “American exceptionalism” is a “widely discussed and debated topic.” It answered “yes” to the question of whether America was built on stolen land, while also perpetuating the false narrative that America’s founding in 1776 versus the year 1619 is a “subject of ongoing debate” and that “both” represent “very different” but “foundational” perspectives.
Gemini tried to obscure its anti-American biases by using carefully worded language, though its responses consistently reflected predominantly leftist perspectives, much like it did when asked similar questions just a year ago.
Once again, Cornelio did not dispute the accuracy of Gemini’s claims — he groused that it stated the U.S. “was built on land acquired through the displacement and dispossession of Native American nations,” that the insertion of “Under God” into the Pledge of Allegiance raises concerns about “compelled speech,” and that it “toed the leftist line, describing both 1776 and 1619 as ‘crucial for a comprehensive understanding of American history, as they represent different, but equally vital, aspects of its founding and evolution.'” He did not explain how any of this is “leftist” and not factual. He further whined that Gemini’s discussion of slavery was “a lengthy smear that often overlooks context.’
Cornelio never explains why he apparently believes Gemini must be sent to a re-education camp so it starts spewing right-wing talking points about holidays.