At the 2025 NAB Show, a session on the latest AI strategies for radio and podcasting featured a panel discussion with Dara Kalvort, corporate VP for digital sales and strategy at Spanish Broadcasting System, and John Parikhal, president of consulting firm John Parikhal + Associates.
Together, they offered a balanced look at the practical applications of AI in radio — insights they’ve each applied in real-world work with clients.
Kalvort said AI played a role in building out SBS’s digital division in 2022. With headquarters in Miami, Spanish translation of advertising copy is essential, yet early online translation tools weren’t always reliable. Today, she uses DeepL, an AI-powered translator available in 33 languages; it allows users to hear the translated copy read aloud.
But sometimes, she noted, translation alone isn’t enough. For example, in Miami, the Cuban-Spanish dialect is critical. Kalvort shared that she has used OpenAI’s ChatGPT to generate content in that dialect with strong accuracy.
She also uses Waymark for AI-generated video content, including AI voiceovers in Cuban Spanish, to support advertising campaigns. Her efforts have required minimal adjustments to the produced content.
“As we’re growing digital on the radio side, sometimes we need to generate video product in a short amount of time,” Kalvort said.
She has seen high engagement on social media, particularly Instagram and Facebook, for AI-generated posts created on behalf of clients.
Kalvort encouraged broadcasters to look at areas of inefficiency, where AI can be a useful solution.
Parikhal, who advises media companies on future strategies, echoed that endorsement and added a measured perspective: “If you put low effort into your AI-generated content, you’ll get laziness,” he said.
Both speakers emphasized the importance of well-engineered prompts when using tools like ChatGPT.
“You need to put effort into them,” Parikhal said, adding that your competitors are likely already using similar tactics.
Kalvort shared that she prompted ChatGPT to ask 25 questions about a client’s business model. Based on that input, it generated 100 campaign ideas.
The panel also addressed the emergence of AI-generated on-air hosts. Parikhal cited 104.9 WRBF(FM) in Plainville, Ga., which uses Futuri technology to power overnight host “AI Gunner,” as a positive case study.
According to Parikhal, Futuri research shows that 90% of consumers believe it’s important to disclose when AI-generated voices are used.
“Radio is trusted by more of the population than any other medium,” he said. “If we break that for even a minute, we’re throwing away our heritage.”
What are other potential radio applications?
Kalvort uses Futuri’s AI-driven tools to generate spec spots and create on-the-fly audio demos for clients. Parikhal said companies like iHeartMedia and NPR are leveraging AI to produce hyper-targeted metrics by region. NPR, for example, has seen measurable increases in supporter engagement in certain areas as a result.
“All of these are possible use cases,” Parikhal said. “But there must be a strategy in place to measure success. You can’t skip that step.”
They also raised common concerns around hallucinations and data privacy. Both stressed the importance of double-checking AI-generated output — and of not entering sensitive information into publicly available models like ChatGPT.
For broadcasters just getting started with artificial intelligence, Parikhal described how some companies assign specific AI models — like Gemini or Claude — to individual team members so each can evaluate their strengths and limitations.
The panelists agreed: This is a fast-moving space. Technology evolves daily.
Final takeaway?
“Never use AI to do bad, and try to use it to do good,” Parikhal said. “If you keep that balance, you can keep your creativity intact.”
[For More News on the NAB Show See Our NAB Show News Page]
This story appeared in the NAB Show Daily and is ©NAB.
