Artificial intelligence is reshaping journalism in ways that are forcing reporters, editors and media organizations to rethink how they work, says LSU assistant professor Nihar Sreepada.
Sreepada, who researches AI across public relations, journalism, advertising and political communication, said newsrooms are learning to treat AI as a tool. A tool that can make reporting faster, but still requires human oversight.
“AI, by itself, is not a negative or positive tool,” Sreepada said. “It has benefits if you use it in the right way. Something we need to know is that we cannot escape it. In my research, I call this the ‘cautious embrace.’ We’re trying to understand what tools to use, how best to use them and when to stop.”
LSU announced last December that it had joined the AI Alliance, a global collaboration of universities, research institutions and industry partners working on safe and responsible AI deployment.
Alongside the alliance, LSU’s website notes that more than 180 faculty members are engaged with AI research supported by $7 million in funded projects over the last five years. That includes the launch of MikeGPT, an LSU-specific AI assistant and a partnership between LSU and Data² to tackle cybersecurity, energy and infrastructure challenges.
“Just like you use Google or social media, AI is another tool that can help with a lot of things,” Sreepada said. “We either overhype it or underrate it, but it’s still a tool.”
With this momentum, Sreepada’s remarks carry a new urgency: Students now studying journalism may soon enter newsrooms that already use AI-driven tools for monitoring, transcription and pattern detection.
“Major outlets like the Associated Press are using AI to generate data-driven stories and basic election results, weather updates and market reports automatically,” Sreepada said.
He explained how tools such as Otter.ai and Google transcription software have already changed how journalists record and process interviews.
While technology can streamline work, it also raises ethical questions.
“Different organizations have different policies,” he said. “In journalism, authenticity has to be there. Public relations specialists might use it more openly, but journalists have to be careful.”
Sreepada shared examples of major newsrooms building their own AI systems. Reuters, for instance, has developed Lynx Insights, a program that detects story patterns across massive data sets, while The Washington Post uses Heliograf, an in-house AI that tracks breaking news and writes short updates.
“Say I want to look at all the cases of human trafficking in the past two to three years,” he said. “Then they can use that and redefine patterns across the past two years. That frees up reporters for more in-depth stories. Before, you used to look for news. Now you can ask for it. That’s journalism being redefined.”
Sreepada demonstrated The Washington Post’s Ask the Post AI, a collaboration between OpenAI that allows readers to ask questions and get summarized answers based on the outlet’s reporting.
LSU officials acknowledge AI is not foolproof. MikeGPT’s documentation states that, although the system strives for accuracy, users should verify all information.
“Editors are treating AI as a starting point, not a final copy,” he said. “It requires human fact-checking and editing and I can’t stress that enough. Many newsrooms and workshops are starting to educate journalists on AI tools and ethics.”
AP has taken a similar stance, directing reporters to treat AI outputs as “unvetted source material” that is useful for speed but never a substitute for human reporting.
“The way I see AI literacy is what tools we use, how to use them and when to stop using them,” Sreepada said.
