A seminar titled “AI: Challenges and Opportunities in Tech Journalism” was held recently at a resort in Bhaluka, Mymensingh, organised jointly by the Center for Technology Journalism (CTJ) and Opus Communications Limited.
Speakers said AI is changing how journalism works. Its use in gathering news, analysing data and producing reports is creating new possibilities, while also creating new challenges for professional skill, accuracy and responsibility. The discussion covered the use of AI-integrated software, data analysis, and applications of AI internationally.
What was said
Hasan Zakir, CTJ’s president, said surviving on traditional methods of journalism is becoming difficult, and that to survive in the digital world journalists must become skilled in AI and the latest technology — and that this skill will bring new opportunities.
Saif Ahmad, CTJ’s general secretary, set out the practical benefits, particularly in data analysis: the technology can reduce the time analysis requires, making journalists’ work faster and more effective.
Md Abdullah Al Kafi, managing director of Opus Communications Limited, presented various aspects of AI-integrated software, and discussed how professionals in mass communication and journalism can use such software effectively as technology changes.
Author and CTJ member Rahitul Islam, drawing on examples from Japan and China, said: “AI has brought new opportunities for us. There is no fear of journalists losing their jobs because of AI; rather, AI has added a new dimension to journalism.”
Kazi Sohag, a CTJ member and special correspondent for Bangladesh Pratidin, said technology is constantly changing and media workers will have to adapt to it, or risk falling behind.
CTJ members took part in day-long activities alongside the seminar. The organisers expect such initiatives to increase knowledge-sharing and contact among technology journalists, and to help professional standards keep pace with technological change.
The agreement is the problem
Every speaker landed in the same place: AI is an opportunity, adapt or fall behind. That is probably true, and it is also the least useful thing a newsroom can conclude. “Become skilled in AI” is an attitude. What a newsroom needs is a document.
The reason is simple arithmetic about risk. If a reporter uses a model to speed up data analysis and it is right ninety-nine times, nothing happens. The hundredth time it invents a figure, a court case, or a quotation, and the paper prints it, the cost is not a correction. It is the credibility of everything else on the page.
The four lines
What may be drafted. Headlines, summaries, translations, structure for a story the reporter already knows — these are low-risk. Anything where the model supplies the facts rather than the phrasing is a different category.
What must be verified, and by whom. Every number, name, date and quotation that passed through a model gets checked against a primary source by a human who is named in the file. Not “we check things” — a named person, logged.
What readers are told. Disclosure rules decided in advance, not improvised after someone asks. If a translation was machine-assisted, say so. If an image is generated, label it in the caption, not in a policy page nobody reads.
What never leaves the building. Unpublished source material pasted into a consumer chatbot has been sent to a third party. For a reporter with a confidential source, that is not a privacy question, it is a safety one.
Two things worth bringing to the next one
First, verification tools now exist on the other side of the problem. Google has opened its AI-detection portal to everyone, which is directly useful to a desk handling submitted photographs — with the large caveat that it only reliably identifies images watermarked by participating generators.
Second, Rahitul Islam’s point that AI will not cost journalists their jobs deserves to be tested rather than asserted. The more likely risk is not replacement but erosion: if the model drafts, summarises and analyses, the reporter gradually stops doing the thing that made them worth employing. That is a real and measurable effect — we wrote about the quiet cost of letting AI do the thinking yesterday — and it hits a newsroom through the junior reporters who never develop the judgement, not through redundancies.
None of which argues against the seminar. Tech journalists in Bangladesh meeting to discuss this is better than not meeting. But the useful output of the next one would be a one-page policy that a newsroom can actually put on a wall.
