How AI and Social Media Are Reshaping Trust in Journalism
Artificial intelligence (AI) has quickly become a part of daily life and the way people create, find and consume their information. News organizations are experimenting with AI to write stories, summarize information and organize content. Even everyday people can produce AI-created news to post for anyone to see. For anyone consuming this content, it may raise a question: If people already question whether traditional media is biased, can they trust AI to provide something more verifiable?
The problem is not simply that AI can make mistakes. AI systems learn from information created by humans, including the news and other material available online. If that information contains patterns of bias, AI can potentially reproduce those patterns in the content it creates. Social media can then spread that content to large audiences, sometimes without the context that would help people evaluate where the information came from.
A recent study by Chengxin Lyu and Fei Wu, published in the Journal of Broadcasting & Electronic Media in July 2026, examined whether AI-generated news is more or less biased than news written by humans. The researchers compared 18,282 AI-generated news articles with 18,210 human-written articles. They examined three types of bias: coverage bias, gatekeeping bias and statement bias.
One of the findings is how AI decides what information to emphasize. The researchers found that AI-generated news focused heavily on higher visibility topics, including international politics, social incidents and armed conflicts. AI-generated news also organized topics into more separated clusters, while naturally written news showed more flow coverage across political, social, cultural and technological subjects.
This is important because bias is not always obvious. Bias does not have to mean that a news story openly takes one side over the other. It can also appear in what gets covered, what receives the most attention and how information is presented. If an AI system repeatedly emphasizes certain events while giving less attention to other, softer news stories, audiences may receive a narrower picture of what is happening in the world.
One explanation for the bias is the information AI learns from. Lyu and Wu explain that generative AI systems are trained on large amounts of natural written text. As the researchers note, AI can “inherit biases from its training data,” including stereotypes and uneven patterns. AI does not begin with a completely neutral understanding of the world; it learns from information that already contains human perspectives.
This creates an important problem for trust.
If audiences already question whether traditional news organizations are biased, AI does not solve that problem. Instead, it introduces another layer of uncertainty. Audiences may now have to wonder whether the information they are reading was written by a journalist, generated by AI, summarized by AI, or reshaped by an algorithm.
Social media adds another layer because users may encounter AI-generated information alongside human journalism and personal opinions.
As a result, audiences have more information than ever but may struggle to determine which information they can actually trust.
This is where Communication Cultivation Theory becomes important.
Cultivation Theory, developed by George Gerbner and his colleagues, focuses on how long term exposure to media messages can shape people's perceptions of reality. The theory does not argue that a television program, article or social media post automatically changes someone's beliefs. Instead, it focuses on the potential effects of repeated exposure to particular patterns of messages.
That idea can be applied to today's digital news environment. Where people may repeatedly encounter news stories, AI-generated summaries/pictures and influencer commentary and not know what is real and what isn't.
A carefully reported story can become a shortened post, AI summary or commentary from someone who did not write the original story. As information moves further from its original source, audiences may have a harder time knowing where it came from or how it was shaped.
Lyu and Wu suggest that AI does not eliminate bias but can reproduce and amplify patterns already found in naturally written news. This matters because journalism relies on human judgment, ethical standards and accountability things AI does not have.
The issue, then, is not simply whether AI can write news. It is whether audiences can understand where their information came from, how it was produced and whether they can trust it. With repeated exposure to AI, social media and claims of media bias, the biggest risk may be that audiences stop trusting journalism altogether.
Works cited Lyu, Chengxin, and Wu, Fei. “Does AI Amplify or Mitigate Media Bias? A Comparative Analysis of Coverage, Gatekeeping, and Statement Bias in AI-Generated and Human-Written News.” Journal of Broadcasting & Electronic Media. Accessed 9/20/2026. Link