Let’s be honest: AI has brought some real benefits to the news industry. It speeds up writing, makes research easier, and lowers costs compared to the past. But behind those advantages lies a big problem: AI can mass-produce fake news—and it’s getting better at making it look real.

It’s actually a little scary when you think about it. A video that seems to show a real government official speaking, or a “certain” policy-change report, might just be stitched together by AI. If people can’t tell what’s real and what’s fake, the credibility of news will be completely destroyed.

At the start of this year, the Global Risks Report from Davos highlighted this issue, naming “misinformation and disinformation” as a major global risk. AI isn’t just an efficiency tool—it’s completely transforming how rumors are manufactured. What used to be like a small-scale workshop run by people writing fake stories by hand has become an assembly line: data-driven, scalable, low-cost, even customizable.

In the past, journalism had real barriers to entry and professional workflows. Reporters had to go on site, do interviews, check documents, verify sources. Editors reviewed everything carefully—dates, data, sources—to get as close to the truth as possible, even if it was slow. That entire system is getting steamrolled by AI.

An AI model can draft a news article in seconds, using smooth language and logical structure that looks perfectly credible. Image-generation tools can create “photos” from just a sentence that look real. Video synthesis tech is even scarier: it can blend someone’s face and voice into a hyper-realistic fake “interview.”

A lot of people wonder: Why is AI so good at making stuff up?

In truth, this isn’t really new. When you scroll through short videos or click on articles, there’s plenty of “clickbait” with deliberately exaggerated thumbnails to rack up views—no one cares if it’s true. AI is the same way. It’s just an obedient tool: it does exactly what you tell it to do. It won’t ask, “Are you sure you want to lie?” If you use it for serious reporting, it can save time and boost productivity. But if you want to make things up, it won’t hesitate.

What’s worse, AI learns from the data it’s given—and that data can itself be false or flawed. A study from NYU found that introducing just 0.001% false information can cause major errors in large language models. The cost of doing that? As little as five dollars. In other words, AI systems are inherently vulnerable to “learning the wrong thing.”

AI also suffers from so-called “hallucinations,” where it makes up content to sound convincing. On top of that, models often lag behind real-time events, and their web search sources vary in quality. All of this leaves a huge opening for generating false content.

At its core, AI isn’t inherently harmful. The real issue lies in how people decide to apply it.

If we want to maintain standards of truth in journalism in the AI era, it won’t be easy. But it isn’t hopeless. Here are a few ideas:

✅ Newsrooms need to upgrade.

The most practical step is to adopt AI detection tools in editorial workflows and build “human-machine” verification systems. AI can help quickly gather data and flag potential issues, but humans still need to make the final call. Combining machine efficiency with human judgment is the best way to stop ultra-realistic fakes.

✅ Train journalists and editors to understand AI.

Journalism schools and media companies need to update their curricula. It’s not enough to teach writing and video production; they need to include AI principles, uses, and risks. Journalists should know how to spot AI-generated fakes and understand the technology’s weaknesses so they don’t get fooled.

✅ Regulation needs to keep pace.

Industry self-regulation isn’t enough. Governments need to set clear rules for publishing AI-generated content and define legal responsibility for producers and distributors. If someone deliberately uses AI to create and spread fake news, there must be accountability and consequences to curb this at the source.

In the end, AI offers journalism greater speed and efficiency but also introduces serious new risks. We can’t afford to dismiss emerging technology out of fear, nor should we blindly embrace it just because it’s popular. How effectively we manage AI depends on our principles and the safeguards we put in place.

If we want journalism to remain credible and respected in the AI age, we need to learn not only how to use AI—but how to control it. Only then can we keep saying, “This is trustworthy news.”