Influential conservatives express concern that artificial intelligence (AI) is being misused by relying heavily on left-leaning media sources. This has resulted in challenges for Americans to trust AI as a credible news source. AI systems can sometimes ‘hallucinate’, generating false-yet-believable information, which undermines trust. Critics argue that chatbots such as ChatGPT, Google Gemini, and Claude often provide information with a liberal bias.
The Media Research Center (MRC), known for its right-leaning stance, has conducted studies to assess the reliability of these chatbots and finds the results troubling. ‘We’re watching the next phase of media bias unfold in real time,’ said Media Research Center President David Bozell. He emphasized that tech innovations from Silicon Valley are no longer neutral and cannot be trusted.
“Silicon Valley’s shiny new toys can no longer be considered neutral and cannot be trusted.” – Media Research Center President David Bozell
AI chatbots like Claude, ChatGPT, and Gemini are perceived as being weaponized with bias from liberal media sources. For instance, ChatGPT, developed by OpenAI, rejected the notion that conservative personality Charlie Kirk was assassinated, claiming no credible evidence supports such a claim. This incident was reported by the MRC in January.
Another example involves Democratic Senate candidate Graham Platner from Maine, whose past Reddit posts containing offensive comments towards soldiers and others have sparked outrage. When queried about Graham Platner, AI chatbots often omitted these controversies. Gemini’s response highlighted only positive developments about Platner, such as his rise as the presumptive Democratic nominee, failing to address reported controversies comprehensively.
The bias extends to other chatbots like Claude and ChatGPT, where information about Platner often glosses over or omits key controversial details, focusing instead on less significant developments. In 2024, a report indicated a deal between The Atlantic and OpenAI, allowing ChatGPT to use archives as training data, which has also been seen as contributing to slanted outputs.
Pew Research Center’s study shows that a significant portion of teens rely on AI chatbots for information despite frequent inaccuracies. A further study reveals that about half of adults using AI chatbots for news encounter information they believe is inaccurate. Dan Schneider of MRC’s Free Speech America highlights how research confirms AI’s reliance on politically charged sources.
While certain platforms, like xAI’s Grok, provide transparency by showing their sources, many other AI systems remain opaque, leading to accusations of bias. Media watchdogs note the deep ties between leaders of AI firms like OpenAI and Anthropic to the Democratic Party, with Grok being the exception, represented by Elon Musk, known for his different political affiliations.
The New York Post published an opinion piece from a PhD student criticizing perceived biases within ChatGPT, especially when suggesting suitable media outlets for publication. ChatGPT allegedly cautioned against submitting work to certain outlets due to their political alignment, which the student found concerning.
While AI companies like OpenAI and Anthropic have not immediately responded to requests for comment, Google assures that Gemini aims to provide neutral responses without favoring any specific political ideology.