AI gives us powerful verification tools but it cannot force us to use them. Technology can help, but judgment decides. The ethical part of journalism remains exactly the same: slow down, question every claim, and check before publishing. Accuracy is still more important than speed.
By Aleksandar Manasiev — NarativAI
As with many young journalists, I began my career in a newspaper in one of those newsroom beats nobody chooses first, the sections where you work harder than everyone else just to prove yourself. But the case I’m about to describe happened in 2013, when I was no longer that young reporter. By then, I already had almost a decade of experience on the crime and investigations desk.
At the time, the Balkans were full of construction scandals, people paying for apartments and receiving nothing more than a piece of paper. Disinformation on social media was still rare back then, especially fabricated stories connected to our region. But “rare” didn’t mean “impossible.”
One day in June 2013, a photo began circulating on Facebook. It was a screenshot from a newspaper somewhere in Latin America. According to the headline, the married couple at the center of one of our biggest scandals had been murdered in that distant country.
Dozens of news portals, and even one television station, published the story as fact. Nobody checked anything. Nobody asked a question. Nobody slowed down.
I was working at a daily newspaper with a 4 p.m. print deadline. I simply googled the name of the foreign outlet, checked where it was based, and sent them an email. Because of the time difference, I didn’t expect a quick answer. It arrived one or two days later, confirming what I had already suspected.
The paper was real, the front page was real but the main story had been replaced. Someone had photoshopped the screenshot to make it look like the couple had been killed, they changed the headline and the main photo. The motive behind the fabrication was easy to guess. But the point is much bigger: if it was difficult to separate truth from fiction in 2013, today the challenge is far greater.
Then VS Now
Back then, verification was slow, often manual, and depended entirely on the reporter’s discipline. If something looked suspicious, you contacted the source, waited for a reply, compared versions, and checked details one by one.
It took time and patience, but it worked, because suspicion pushed you to verify. What protected journalism in 2013 was not advanced technology, it was the decision to pause, question, and not be hasty.
The paradox
Today, we live in a world where verification can happen in seconds. Artificial intelligence can compare visuals, translate foreign content instantly, analyze manipulation patterns, and search across thousands of websites at once. The fake story that spread across the Balkans in 2013 could be debunked today with a single upload or a quick reverse search.
But here is the paradox, the faster verification becomes, the easier it is to skip it. Newsrooms race to be first. Journalists publish under pressure. Social platforms reward speed over accuracy. And when everything moves at the pace of a swipe, suspicion becomes even more important than it was a decade ago.
AI gives us powerful verification tools but it cannot force us to use them. Technology can help, but judgment decides. The ethical part of journalism remains exactly the same: slow down, question every claim, and check before publishing. Accuracy is still more important than speed.
The strongest form of verification today is a blend of technology and skepticism. AI can expose manipulation, but only if the journalist chooses to verify. AI can speed up the process, but it cannot replace responsibility.
The lesson from 2013 still applies in the era of deepfakes and synthetic media: suspicion is not a weakness, it’s your strongest professional reflex. AI can help us find the truth faster, but only journalists can decide to look for it.
(This text was written and reviewed by the editor with support from artificial intelligence tools for language editing and stylistic refinement. More on how NarativAI uses AI — Link)