According to a report by Reuters, Publisher Gannett is planning on introducing generative AI in its systems. But they plan on doing so with caution, provided that humans have the final word on any releases.
The goal is to use the system to publish stories in a bid to both improve efficiency and save money. Gannet is the largest newspaper publisher in the United States with over 200 daily outlets. The company claims that it won’t leave humans out of the process. In short, generative AI systems won’t have the ability to deploy automatically.
In an interview with Reuters. Renn Turiano, senior vice president and head of product at Gannett, said that generative AI is a new method of removing tedious tasks clogging up journalists’ time and providing the groundwork for new efficiencies.
But with that said, he noted that “The desire to go fast was a mistake for some of the other news services,…We’re not making that mistake.” Though he didn’t name a specific news organization, it’s clear that Gannett and others within the news business are very concerned with any mistakes that AI can create.
As you’d imagine, Gannett, isn’t the only news organization looking to embrace generative AI. Reuters President Paul Bascobert said in a statement on of their company’s plans for AI Thursday, “taking a responsible approach that safeguards accuracy and fosters trust.”
Newsrooms and organizations were some of the first hit by AI technology with the explosion of ChatGPT and other Large Language Models. Some, such as CNET, known for its focus on technology and gaming, took the leap into AI quickly. But after pushback from staff, the company backed off early this year.
One thing to note is that according to CNET, they were testing an AI engine and learned quite a bit from how humans and AI can work together. Though CNET’s approach was cautious, and it seems Gannett took note of their own move toward AI, many are still warning of drawbacks.
For example, Nicholas Diakopoulos, Northwestern University associate professor, said of LLMs and their ability to take on journalism work independently. “Where I am right now is I wouldn’t recommend these models for any journalistic use case where you’re publishing automatically to a public channel.”
As of right now, Gannett says that it plans on rolling out a live pilot program using AI to identify the most important parts of an article. Then create a bulleted summary to match. The goal is to launch this program sometime at the end of the year. But, they made clear that journalists will still have the final say when it comes to using any AI-generated work.