As newsrooms increasingly experiment with artificial intelligence, panelists for the LMA Fest session “AI in Action — Innovative Ideas for Media Companies” urged transparency and cultural adaptability. Emphasizing audience trust and ethical standards, the speakers shared strategies for using AI responsibly in journalism.

Dorrine Mendoza, product and partnerships lead at the American Journalism Project’s Product and AI Studio, shared her organization’s approach to supporting newsrooms in AI experimentation. Through a $5 million OpenAI grant, Mendoza’s team has guided newsrooms using AI-driven tools, including a GPT-based election assistant.

Mendoza stressed the need for newsroom cultures that support AI experimentation, particularly when leaders themselves are hesitant about the technology. She noted that successful AI integration works when “everyone in the organization feels empowered to experiment with AI and to report back on what they’re learning.” A flexible AI policy, she explained, allows room for experimentation without fear of repercussions.

Lynn Walsh, assistant director at Trusting News, highlighted her work with newsrooms on transparency in AI. “We help journalists rebuild trust primarily through transparency and engagement,” she explained.

Walsh noted that trust issues in journalism often arise from public assumptions, which are typically negative and frequently incorrect. She stressed the importance of transparency, explaining that readers expect to be informed if AI is used. In her research, 94% of news consumers wanted to know when AI is used in journalistic processes. “I think the message is clear, right?” Walsh said. “They want disclosure.”

Mike Reilley, a senior lecturer at the University of Illinois Chicago and founder of Journaliststoolbox.ai, underscored the value of AI for efficiency in journalistic workflows but warned against over-reliance on automated tools. “No technology can live up to the hype of AI,” Reilley remarked, but find one or two things you need AI to do, start there, and then grow from it.

Reilley also urged newsrooms to create policies, such as those outlining freelancers’ use of AI, to prevent poorly crafted content generated by AI usage.

Audience questions addressed various concerns, including transparency in AI disclosures and the public’s trust in AI-generated content. Walsh responded by advocating for clear, accessible disclosures emphasizing AI’s benefits in reporting, such as enabling journalists to cover more stories efficiently. Walsh suggested an evolving feedback loop with the audience to ensure transparency remains relevant as technology advances.

Reilley and Mendoza discussed practical steps for newsroom leaders to familiarize their teams with AI. Mendoza suggested leaders experiment with AI for everyday tasks to understand its potential and limitations. Reilley recommended that newsrooms test multiple AI tools to find the best fit for specific tasks.

The session highlighted practical advice for small newsrooms. Walsh recommended starting with a community survey to gauge audience comfort with AI in reporting, while Mendoza suggested simple team conversations to understand employees’ current AI use. Reilley added that many technology companies are open to partnerships and could provide AI tools for journalists at reduced costs or free trials.

Key takeaways

  • Transparency builds trust: Walsh emphasized that clear, honest disclosure of AI use can strengthen reader trust. Research shows that 94% of news consumers want to know when AI is involved in creating or editing news content.
  • AI needs flexible policies: Mendoza highlighted the importance of flexible AI policies within newsrooms, which allow for experimentation without penalizing users. She noted that successful AI integration depends on a supportive culture that includes leadership buy-in.
  • Detection tools for deepfakes: Reilley highlighted tools like Deepfake-o-meter that detect AI-generated images and videos. He argued that such tools are essential for journalists to verify content and prevent misinformation from spreading.
  • Experimenting with multiple AI tools: Reilley advised journalists to test several AI tools to determine which are best for different tasks.
  • Simple AI start for small newsrooms: For newsrooms new to AI, Mendoza and Walsh suggested small-scale experiments and community surveys to gauge public comfort, promoting incremental and audience-aligned integration of AI tools.

Editor’s note: Artificial Intelligence was used to transcribe and create an initial summary of this article edited by LMA staff.