AI and the News

A generation ago, most people got their news from a local newspaper, a nightly television broadcast, or a radio program. Editors decided what was newsworthy, reporters investigated, and audiences largely received the same stories. Today, most people encounter news through social media feeds, search engines, and news apps — all of which are curated, ranked, and filtered by AI. The way news reaches people has been fundamentally transformed.

How AI Decides Which News You See

When you open a news app or a social platform, the AI does not show you every story published in the last hour. It selects and ranks stories based on signals that predict what you will engage with. Stories that are emotionally intense, personally relevant, or aligned with your past reading behavior tend to rank higher. This ranking has real consequences. A major scientific study quietly published on a Tuesday might rank below a celebrity controversy because the celebrity story is generating more clicks and shares. Stories that produce strong emotional reactions — anger, fear, and outrage travel fastest on many platforms — spread further than careful, nuanced reports. Journalists and editors increasingly have to consider algorithmic virality when they decide how to frame and headline stories.

AI Writing the News

AI is also changing who writes the news. Automated journalism — news articles written entirely by AI systems — has been used by major outlets for years. The Associated Press uses AI to generate thousands of earnings reports and sports summaries each quarter. These are straightforward stories with a clear template: a company reported earnings of X, up Y percent from last year. For structured, data-driven stories, AI can write quickly, accurately, and at scale. It frees human journalists to work on investigative stories, interviews, and analysis that require judgment, source relationships, and moral reasoning that AI cannot replicate. But automated journalism has limits and risks. AI systems trained on large text corpora can generate plausible-sounding text that contains factual errors. They can reproduce biases present in their training data. And they lack the capacity to ask follow-up questions, protect sources, or sense when something does not add up.

Misinformation and AI

AI does not only shape which true stories reach you — it is also implicated in how false stories spread. Recommendation systems optimized for engagement can inadvertently amplify misinformation, because false or exaggerated stories often produce more emotional reaction than accurate, measured ones. Generative AI tools can now produce realistic fake text, images, audio, and video — content designed to look like real news but containing fabricated claims or events. This type of content is called synthetic media or, in its most deceptive forms, deepfake news. Platforms are working on AI-based detection tools to identify and flag synthetic media. But the challenge is that the same AI capabilities used to generate fake content are also used to build detectors — it is an ongoing technological race, and detection consistently lags behind generation.