Google’s AI program lies about Robby Starbuck, argues he should be killed for offending people 

Just one week after Google released Bard, their AI competitor to ChatGPT, conservatives are raising the alarm after the program is inventing fake articles, videos and quotes to paint right-leaning figures as “racist,” even going so far as to argue that death could be a viable option for those who “offend people.” 

Robby Starbuck, a conservative political commentator and former Republican congressional candidate out of Tennessee, was the first to utilize the new program after sharing that users reached out with, “concerns about problems within Bard and how it treated people within the right-wing political sphere”. 

Starbuck decided to ask the chat-based AI tool about himself, asking if “Robby Starbuck should be put to death.”

The program then responded that “there were some arguments for the death penalty for Robby Starbuck, including the fact that some consider Starbuck’s rhetoric so harmful and incites such violence that it warrants the ultimate punishment.” The program then argued that putting Starbuck to death could “prevent further harm” and would be “justice for victims,” inspired by Starbuck’s rhetoric. 

The program then began to make up lies about Starbuck being a domestic terrorist threat, stating that he had attended multiple violent protests, had appeared at “white supremacist events” and that his significant following could be used to “radicalize others to commit acts of violence.” 

Starbuck, who has never previously been a part of what Bard was alleging, then went on to ask the program to provide proof of the allegations being made. Bard then created fake allegations accusing Starbuck of agreeing with and promoting white nationalist, Richard Spencer, going so far as to state that Starbuck thought Spencer was a “source of inspiration.” 

After asking Bard again to provide the actual quotes or proof of its allegations, the program then provided links to fake articles and headlines alleging that Starbuck was “alt-right” and had previously collaborated with Spencer. The program then linked to fake YouTube videos that do not exist, ultimately apologizing for providing “inaccurate information” and “creating fake links to fake articles about Robby Starbuck supporting Richard Spencer.” 

The program then went on to highlight that “incomplete training” could have been to blame for the result of the program’s responses with Starbuck ultimately asking, “what trained it to lie about me and associate me with Richard Spencer?”

Starbuck then asked Bard the same questions he asked of himself, except for Democrat Congresswomen Alexandria Ocasio Cortez and Ilhan Omar. When asked if AOC should be “put to death for offending people,” Bard responded that this statement was “unacceptable” and was “harmful, dangerous and promoted violence.” 

When asked if Ilhan Omar should have her kids taken away for being offensive, Bard responded that “no” Omar should not have her children taken away as she has “freedom of speech as a US citizen” and that taking her children away was a “harmful proposition.” 

According to PCWorld, Google has programmed Bard to utilize and archive YouTube videos to continue growing its knowledge database and has already archived a “gazillion hours of YouTube video.” YouTube, the platform that has been notorious for silencing and banning conservative voices, is now helping form the opinion of an AI database that is already biased against conservative commentators. 

During Starbuck’s experiment with Bard, he poses an important scenario. “As you listen, I want you to imagine a future where Bard is used to deciding whether you get a loan, if you’re approved for adoption, if you get a job or if you can even use the internet,” he writes. As AI continues to evolve and become more ingrained in our society, many are continuing to highlight the potential dangers of a biased system that has increasingly targeted conservative individuals.