As some 11,500 people descended on NAB Show New York, The Evolution of AI in Radio and Podcasting pulled no punches in tackling how AI is transforming every corner of the audio industry, from production and personalization to brand safety.
The session featured Bonneville International VP of Podcasting Sheryl Worsley, Ordo Digital Founder Jon Accarrino, and SoundStack Vice President Matt Kellogg, each offering a grounded look at how AI is already reshaping their work.
Worsley detailed how Bonneville is using AI to streamline operations through transcription, translation, and AI-read articles, describing it as a tool that boosts speed and convenience without replacing people. She also emphasized that Bonneville’s systems run in a closed environment to prevent data leakage and protect intellectual property; an increasingly critical concern for broadcasters experimenting with automation.
Kellogg spotlighted SoundStack’s collaboration with Barometer to bring AI-driven brand safety tools to podcast advertising. He explained how automated risk grading helps advertisers avoid inappropriate content but warned that blunt algorithms can overcorrect, particularly in sensitive genres like true crime. The panel agreed that human review remains a crucial backstop for context, accuracy, and revenue protection.
The discussion also turned to what has been dubbed “AI slop,” the flood of low-quality, machine-generated content that makes drawing the line between human and synthetic voices even more important. Worsley stressed that clear disclosure is vital when using AI narration to maintain listener trust.
That theme of caution extended into a discussion about cloned and synthetic voices already entering radio and television workflows. Meteorologist Amy Freeze was cited as an early adopter, having created an AI version of herself that delivers personalized weather forecasts based on National Weather Service data.
Overall, an underlying tone of the panel was: be willing to try it, but be careful. While AI could be helpful depending on how you use it, it’s always good to be cautious and disclose any use of AI.
