Memory and AI.

 

Radio stations have thousands of hours of archives, often underutilized because they are difficult to index, browse or restore. AI can transform this dormant memory into an active resource, harnessing transcription, keyword searching, automatic summary and thematic upgrading. When direct reporting is impossible, coverage can be enhanced by historical archives. This idea raises the question of what we preserve, what we transmit, and the editorial mission linked to how we use augmented memory.

Useful links:

Memory of the World
Managing low-cost digitization projects in Least Developed Countries and Small Island Developing States



Sub-themes around this idea for programme content: ways to leverage, develop or angle the topic:


For listeners:

• Testimonial: rebroadcasting or reinterpreting old feature segments
• Memory engine or erasure machine?
• Personalized access to sound archives, on air in real time
• Personalized access to sound archives, via interface
• Participatory features: “Tell us which radio memories you want to listen back to”
• Live voting

Potential guests: radio archivist, documentalist, AI audio engineer, university chairs, specialists in collections of humanity's shared heritage, etc.


For the radio teams:

• How can we use AI to add value to archives?
• AI for summarising and restoring old content
• Voice search engines in older podcasts
• 10 years of podcasts in 10 minutes
• Using archives to hold executives to account
• Instant real-time contextualization during a debate or interview