Concerns Over AI Bias in Media Detection
Many influential conservatives are expressing their belief that artificial intelligence is heavily influenced by left-leaning media, which, in turn, makes it hard for the average American to trust these AI-generated news sources.
It’s pretty clear that AI systems have a tendency to “hallucinate,” meaning they can generate false yet convincing information, which could potentially undermine public confidence. Some skeptics argue that chatbots like ChatGPT, Google Gemini, and Claude skew their information towards liberal sources, effectively overshadowing opposing viewpoints.
As more individuals turn to these technologies for information, the Right-Wing Media Research Center (MRC) undertook a study to assess the reliability of chatbots. The results—they weren’t exactly promising.
David Bozell, the Director of the Media Research Center, remarked, “We are witnessing the next phase of media bias in real time. Silicon Valley’s new toys have lost their neutrality and can no longer be trusted.”
In January, the MRC pointed out an instance where ChatGPT, created by OpenAI, incorrectly stated that no one named Charlie Kirk was assassinated, claiming there was no credible evidence to support it. The MRC also pointed out that Claude, developed by Anthropic, rejected the notion that AI should factor in First Amendment principles in its operations.
In a related but disturbing development, Fox News Digital recently reported that Graham Platner, a Democratic Senate candidate from Maine, faced backlash over crude comments he made on a now-deleted Reddit account. His posts included graphic remarks regarding inappropriate behavior in portable toilets and a mocking attitude toward a soldier who suffered serious injuries in combat.
This Reddit controversy has sparked considerable outrage, especially his disparaging comments about injured soldiers, suggesting that they didn’t deserve to live.
Despite the heightened controversy surrounding Platner, major chatbots downplayed the issue when they were asked about him. Gemini mentioned that Platner has significant backing, noting his past cover appearance on Time magazine and his polling lead against Sen. Susan Collins.
Gemini did acknowledge that Platner faces scrutiny over his past online presence but primarily focused on positive developments regarding his campaign. In contrast, ChatGPT highlighted how the attacks from Republicans were making the old Reddit posts central to Platner’s narrative.
In dealing with this particular incident, ChatGPT referenced a Bangor Daily News article detailing how Republicans are dissecting Platner’s past on Reddit, but access to the original material is hindered by paywalls around Maine’s local newspapers.
Claude offered some summary on Platner’s political journey and rising popularity, even touching on a controversial Reddit comment where he hinted at a communist past, stating it was a joke but neglecting to mention the controversy surrounding his recent comments.
Meanwhile, Nicholas Thompson, CEO of Atlantic, emphasized the need for shaping the industry based on certain values, which could sway how news is displayed across platforms, potentially favoring established brands over alternatives.
A Pew Research Center survey revealed that many teenagers rely on chatbots for information, with a significant portion believing the news they encounter can be inaccurate—from false narratives to partisan slants. Almost half of adults reported encountering what they perceive as inaccurate news from AI chatbots.
Dan Schneider, a vice president at MRC, attributed the bias in today’s AI systems to a predominant reliance on left-wing media for generating content. He argued that whereas some platforms are beginning to show sources for their claims, the overarching industry is still opaque.
Media analysts have noted that key players in the leadership behind AI platforms like OpenAI and Anthropic hold significant ties to the Democratic Party, contrasting sharply with Grok, which is associated with Trump ally Elon Musk.
In other shocking revelations, a Cambridge University student shared her experience with ChatGPT, noting its bias when recommending where to publish her article critical of a progressive journal. Even though the article suited a publication like the New York Post, the chatbot cautioned that it would potentially harm her career.
Marks found it unsettling that publishing in what she deemed a major outlet could negatively reflect on her professional credibility purely based on political orientation.
Interestingly, ChatGPT suggested she tailor her piece to fit left-leaning publications instead. Responses from OpenAI and Anthropic were not immediately available for comment, while Google affirmed that Gemini is programmed to provide neutral responses free from ideological biases.
