This fall we’re starting to see the release of many of the AI tools that have been teased all year.
Meta.AI, created by the parent company of Facebook and Instagram, has upgraded its engine and offers robust text and image generation. Meta now also offers AI-aided image editing and a voice-mode using Hollywood likenesses such as Kristen Bell, Awkwafina, Dame Judi Dench, and John Cena.
Down the street in Cupertino, the release of new iPhones and iOS 18 means Apple finally can show off the first tools in its Apple Intelligence portfolio, which also includes AI-image editing.
But the biggest “Oh Wow!” moment in AI this fall has come from Google’s AI labs with a new feature in a product called NotebookLM. This product has been out for a while and was originally designed for helping students organize class materials. Students can upload materials such as class notes, YouTube videos, PowerPoints, lecture recordings, and texts, and NotebookLM will generate cohesive study guides and Q&A sessions to help students learn complex material.
In September, an on-demand podcast generator was added to NotebookLM. Students upload the material, and in addition to the original study material, they can click a button. A two-voiced WAV file is then created with a male and female voice discussing and summarizing all the material that’s been uploaded.
Not only is the breadth of the content discussed in the podcast notable, but what really is remarkable is the humanness of the conversation. Google’s engineers have made sure that the voices sometimes are talking on top of each other, inject the occasional “um,” and use modern-day colloquialisms. It’s a big step from the current audio-generating tools that output audio with the occasional incorrect pronunciations and wrong vocal inflections.
At Jacobs Media, we uploaded a PDF of the 2024 Techsurvey to the tool as an experiment. The output was a lively, pro-radio discussion summarizing what the 30,000 listeners we surveyed were thinking about radio. (Listen to it here: https://bit.ly/40epb7B)
The generative audio in NotebookLM is a hit. There are now over 500 podcasts using this tool in current podcast directories, according to podcast specialist James Cridland, and Google recently announced new features would be added to the generative-audio tool, including tweaks for the business community.
NotebookLM isn’t ready to summarize your radio station’s morning show or your TV station’s six o-clock news. But it’s not hard to imagine in the near future that this tool could use a cloned voice (with permission) to easily create content with a station’s brand on multiple platforms. NotebookLM shows where AI tools are headed, with the ability to expand brands to new platforms with minimal effort.
The WBA Digital Hotline is a free service of the Wisconsin Association of Broadcasters and is available to all members – contact Chris Brunt at Jacobs Media at Chris@jacobsmedia.com with any digital questions you have.
Download Jacobs Media’s free weekly newsletter on AI – The AI Edge – at Jacobs Media and AI (https://jacobsmedia.com/jacobs-media-ai-edge/)
