This American Life 870: My Other Self
15 snips
Oct 12, 2025 
Evan Ratliff, a journalist and storyteller, dives into the fascinating world of AI by creating a voice clone of himself. He shares his wild experiments, from testing it on customer service calls to using it against telemarketers, and even sending it to Zoom meetings. Ratliff reflects on moments when people mistake the clone for the real him, revealing insights about trust in technology. The discussion also touches on the emotional implications of robotic replicas on relationships, making listeners ponder the cost of alternate selves.
AI Snips
Chapters
Transcript
Episode notes
Glitches In First Tests
- Early tests with customer service calls revealed the clone made up account details and often faltered under verification questions.
- The clone sometimes kept talking after hangups or adopted the other party's perspective unexpectedly.
Use Purpose-Built Agent Platforms
- Use dedicated AI agent platforms to build voice agents instead of ad-hoc hacks for more reliable performance.
- Configure instructions like "never take on the role of the person you're calling" before deployment.
Turning The Tables On Scammers
- Ratliff deployed clones against telemarketers and scammers by seeding a phone number into online forms and watched the calls flood in.
- Scammers quickly guessed something was off and sometimes labeled the caller a robot during conversations.