AI for journalism: We’re in the first inning

A handful of monumental technological developments have ushered in marked change in how people both create journalism and access news — desktop publishing, the World Wide Web, the iPhone, and the rise of social media. 

Generative AI is the next breakthrough that will shape our information ecosystem, said Trei Brundrett, the cofounder and former COO of Vox Media who is now advising the American Journalism Project and Aspen Digital on AI initiatives. 

“They broke down walls and gave power for creativity. People are using AI in the same way,” he said speaking at the 2024 Local News Summit hosted by The Lenfest Institute for Journalism and Aspen Digital. “Empowerment creates radical transformation of user behavior.”

Generative AI is still very much in its early days — think of it as still the first inning of the game — and there’s an opportunity for the news industry to shape the technology proactively. 

“We have not been disintermediated by AI…yet,” Brundrett said. There are still a lot of scars from the social era. AI is not just one tech company or platform. It’s a whole new technology.”

There is an opportunity for publishers to develop their own large language models and leverage AI to support their work and revenue models. He acknowledged that many questions still remain, but the best way to answer them is to “work in the open and work together to create new things.” 

Here are some of the strategies the news industry is pursuing and opportunities for collaboration to move forward: 

• While there is rightfully concern about how artificial intelligence can be used for mis-and-disinformation, Claire Leibowicz, had of the AI and Media Integrity program at Partnership on AI, noted that “AI can touch every level of information production and consumption,” highlighting an example of how deep fakes could be used to protect anonymous sources. 
• As publishers experiment with public-facing AI tools, they should ensure that the experiences are:
  • Conversational – how can outlets use chat to unlock content and create unique experiences? 
  • Accessible – most of American journalism happens in English. But AI opens up many opportunities for translation to reach minority language communities?
  • Relevant – The output must be accurate and service-oriented. The product experience should be paramount. 
  • Transparent – Publishers should share what they’re learning in real-time and use this moment of experimentation to educate the public. 
• AI has already shown the potential to expedite and assist with reporting through tools that assist with transcription and coverage of state government from outlets like CalMatters and Spotlight PA. There is ample space for continued experimentation, but industry leaders must agree that AI should not replace journalists, but rather is a tool to help supplement their work. 
• The questions around intellectual property are ongoing with The New York Times’ litigation against OpenAI and many major news sites blocking their sites from being crawled by AI bots. While some news organizations are looking to reach agreements with the major AI providers, there’s also an opportunity for news organizations to leverage their content to build their own large language models. 
• As the AI field evolves, there is a continued need to develop industry-wide standards and ethical guidelines to shape how news organizations leverage AI and present AI work to their audiences. For example, how should news organizations handle user generated content now that both Google and Samsung produce phones that use AI to edit photos on the devices? There are a wide range of questions that need to be answered and the industry should convene groups to set standards and learn together.