On Friday, May 2nd, I attended the NESPA conference at Boston University. There were many great speakers and workshops there, where I learned many great pointers for journalism. I would say the most memorable event I attended was my 1st workshop: AI Agents in the Newsroom.
3 students from Emerson College presented their findings about using an AI agent to assist them with their article. Throughout their presentation, they explained how it helped them, what they learned, and how it can be properly used in the world of journalism. By the end, I was convinced about AI journalism agents.
More importantly, one might ask, “What is an AI Agent?” As described in the workshop, an AI agent is an automated program that can be programmed to complete specific tasks. They’re versatile and are generally accurate in the information they handle.
AI Agents are mostly reserved for doing the slower and more time-consuming tasks in journalism, fact-checking, data gathering, and formatting. They can also handle publishing and notify you after posting or if someone interacts with your post (much like a human assistant could do). They can also cross-check each other’s work to verify the integrity of their statements.
With enough technical knowledge, any aspiring journalist can make their own AI assistant. The students mentioned 2 platforms for building AIs, which were CURSOR and Zapier. Both are publicly accessible, so one could tailor their own assistants to fit their needs. If it were me, I’d call my AI Jarvis, or perhaps WALL-E.
Jokes aside, AI assistants can be useful. In my time so far in The Oakmonitor, I haven’t worked with an AI Assistant to help me check or collect information. SNO sites does not offer any AI features in 2025.
As said earlier, there isn’t much between me and getting an assistant. Both Zapier and CURSOR offer limited but free plans. Their paid plans offer many more features, but are intended for use by large corporations, rather than by a use of 1 journalist.
While this is a great advancement in both the fields of machine learning and journalism. However, there are some underlying issues with AI mixing with journalism.
A prime example of this would be the dead internet theory. Put simply, the theory suggests that the internet is losing human interactions, creativity, and charm, being replaced by AI content creation and interaction. In other words, the internet is gradually becoming more controlled and influenced by AI.
This theory is one of the more extreme ones out there. While it’s somewhat exaggerated, the point is that AI could be replacing humans.
Some people who attended the NESPA AI seminar questioned the ethics of using AI in journalism. In more detail, they asked if the articles would lose their human touch if written with help from the AI agents.
Pew Research, a critically acclaimed fact-tank, conducted some surveys on AI in 2024. The results of one survey showed that 59% of Americans surveyed said AI would lead to fewer jobs for journalists, while only 5% said it would create more jobs. In another survey, 41% expressed concern about AI reporting poorly. Perhaps most importantly, 66% of people were concerned about AI spreading false information and fake news.
Condensing all this, we can take away that most people would prefer human journalism over AI journalism. Most worry that it will take human jobs in the future, and impact the integrity of whatever is being reported.
Yet, there’s a fine line between properly using AI and overusing it. For example, using an AI agent to check the grammar and proper spelling of an article you wrote is fine. While I have not done that yet, I wouldn’t hold anything against somebody doing it. Automating an agent to publish your articles, so long as they don’t edit the content of your piece, isn’t a problem. Using an AI to check your sources is acceptable if done properly.
The problem starts when AI is the writer, editor, source, expert, and publisher for articles. When there is 0 human input, other than the prompt given to the AI, people have the right to be concerned. Both AI and humans can hold political bias, but humans cannot mass-produce articles like AI can.
What can be taken away from all this is that AI and journalism can go together, but only in a delicate balance. Making an AI more like an assistant, such as the AI agents from Emerson College, can prove to be beneficial to a school or college paper. Inversely, giving the AI more power can be detrimental to a paper, especially with overusage.
But to put a positive spin on things, AI assistants and agents make room for experimentation, and can push the envelope for what cutting-edge things AI can do to assist the future of journalists.