The newsroom is no longer just a space for reporters and editors. Today, algorithms sit quietly in the background suggesting headlines, generating content, analyzing trends, and shaping what audiences see.
Artificial intelligence is not the future of journalism; it is already part of its present. But the real question is: are media professionals ready for it?
One thing has become clear with discussions among media professionals these days -- while AI tools are advancing rapidly, AI literacy within media organizations is still far behind. That gap is not just technical; it is editorial. It is ethical and it directly affects public trust.
AI is powerful but not reliable. Globally, news organizations like Reuters and the Associated Press are already using AI for automated reporting and data-driven journalism. In Pakistan, adoption may still be gradual, but AI tools are increasingly being used for content creation, research, and digital distribution.
However, there is a critical misconception: that AI-generated content is inherently accurate. AI can produce information that sounds convincing but is false. This phenomenon is often called hallucination. AI can fabricate quotes, misattribute data, or even invent sources, all while maintaining a confident and authoritative tone. For a journalist working under pressure, this creates a serious risk. Publishing unverified AI-generated content is not just a minor error; it can damage credibility, misinform audiences, and erode trust in media institutions.
Headlines is another affected aspect. Some journalists have admitted that they use AI to draft or rewrite headlines, mostly to save time. Nobody argued that the tools should be abandoned, but there was a shared unease about what happens when a headline is optimized purely for clicks.
AI does not know what a story means; it only knows what tends to get attention, and left on its own, it will usually choose the version that exaggerates rather than the version that is accurate. A journalist recently described a headline his newsroom had to retract because it promised something the actual report never delivered. The takeaway was clear: AI can suggest a headline, but someone still has to check whether it is true before it is published.
Pakistan's newsrooms carry a specific version of this problem. Resources are limited, training varies wildly from one outlet to another, and the pressure to publish quickly usually outweighs the caution required to publish carefully. Most local media professionals have no written policy on whether or how AI should be used in reporting, leaving individual reporters to make these decisions on their own -- often on deadline and often without fully understanding how the tool works.
There is a transparency issue as well: audiences have almost no way of knowing when a story, a photo, or even a headline has been assisted by AI. In a media environment already struggling to maintain public trust, that is not a small blind spot.
Traditional verification isn't enough anymore. Journalists have to treat AI output as a starting point, not a finished fact. Every quote, claim, and statistic it produces requires independent checking cross-referencing primary sources, running visuals through reverse image search, using fact-checking platforms, and staying alert to content that reads a little too smoothly. AI doesn't understand truth; it predicts language based on patterns. As one expert recently put it, "AI can assist journalism, but it cannot replace journalistic judgment."
One alarming figure has been shocking: Apparently, only around 3% of media houses in Pakistan have any formal process for verifying content. Given how much unverified material moves through newsrooms daily, that's a difficult number to absorb.
For instance, for photos, there is TinEye and basic reverse image search, which help trace where an image actually came from and whether it is being recycled from an unrelated, older event. For claims about location, cross-checking against Google Maps can confirm or contradict whether footage was shot where it is claimed to have been. Even the weather can give something away -- tools like Wolfram Alpha can check whether the conditions visible in a clip match historical records for that date and place, a quick way to catch old footage being passed off as current.
For video specifically, InVID is a starting point, alongside basic digital forensic checks: breaking footage into frames, reverse-searching those frames individually, examining whatever metadata survives. None of it requires technical training, but mostly, none of it is taught in journalism programs or newsrooms locally.
Pakistan's media landscape is already vulnerable to misinformation, especially during elections, crises, and natural disasters. Social media accelerates the spread of unverified content, often outpacing corrections. AI has the potential to make this problem worse. Deepfakes, synthetic media, and automated disinformation campaigns can blur the line between reality and fabrication. Without proper AI literacy, journalists risk becoming unintentional amplifiers of false narratives. This makes training and awareness not just important, but urgent.
Media organizations must take responsibility. AI literacy cannot be left to individual effort alone; it must become part of institutional policy. Newsrooms need to introduce structured AI training for journalists, develop editorial guidelines for AI usage and encourage ethical and critical engagement with technology. At the same time, journalists must be equipped to understand how algorithms influence visibility, what trends, what gets suppressed, and how public discourse is shaped online. In today's digital environment, free expression is increasingly mediated by technology.
In the end, being a journalist is no longer just gathering and reporting information anymore; it now means navigating systems that shape how information is created and consumed in the first place. AI is not inherently harmful, but trusting it without question can be. Its place in media is already settled; it is here. What is not settled is how it gets used, and that is what will define the next stretch of journalism. Media professionals need to adapt, not by grabbing every new tool, but by actually understanding what those tools actually do. Because in the end, journalism runs on trust. And in the age of AI, that trust will depend on knowing when not to trust the machine.
The writer is associated with Samaa TV as a producer and researcher
