cross-posted from: https://lemmy.ca/post/28090788

In June 2023, Paul Skye Lehrman and his partner Linnea Sage were driving near their home in New York City, listening to a podcast about the ongoing strikes in Hollywood and how artificial intelligence (AI) could affect the industry.

The episode was of interest because the couple are voice-over performers and - like many other creatives - fear that human-sounding voice generators could soon be used to replace them.

This particular podcast had a unique hook – they interviewed an AI-powered chat bot, equipped with text-to-speech software, to ask how it thought the use of AI would affect jobs in Hollywood.

But, when it spoke, it sounded just like Mr Lehrman.

That night they spent hours online, searching for clues until they came across the site of text-to-speech platform Lovo. Once there, Ms Sage said she found a copy of her voice as well.

They have now filed a lawsuit against Lovo. The firm has not yet responded to that or the BBC’s requests for comment.

  • Eheran@lemmy.world
    link
    fedilink
    arrow-up
    3
    arrow-down
    3
    ·
    2 months ago

    So does it actually sound like them or is this like last time with something like this from that celebrity woman where it was in no way like her?

    • Viking_Hippie@lemmy.worldOP
      link
      fedilink
      arrow-up
      18
      arrow-down
      1
      ·
      2 months ago

      First, in 2019, Ms Sage says a user reached out asking for her to record dozens of generic sounding test radio scripts.

      Test recordings are often used in film and television for focus groups, internal meetings, or as placeholders for works in progress. Because they won’t be shared broadly, these recordings cost much less than audio meant for broadcast.

      Ms Sage says she completed the job, delivered the files, and was paid $400 (£303).

      About six months later, Mr Lehrman says he got a similar request to record dozens of generic sounding radio ads.

      So yeah, seems like it very much sounds like them because it IS them, stolen from other recordings