This American Life

870: My Other Self

October 12, 2025

Key Takeaways Copied to clipboard!

  • Host Ira Glass realized that people, including podcast host Pablo Torre, often construct and interact with an edited, fictionalized version of a public figure based on their media presence. 
  • Reporter Evan Ratliff experimented with creating and deploying an AI clone of his voice and persona, finding that while it could fool some, it often failed in complex interactions, yet it caused genuine distress when used on close friends. 
  • The resistance to new technology, exemplified by a 1924 article about a shopkeeper refusing a telephone, mirrors contemporary anxieties about AI, suggesting a recurring human need to define a 'point of reserve and refusal' against overwhelming change. 

Segments

Ira Glass’s Alter Ego Surprise
Copied to clipboard!
(00:00:39)
  • Key Takeaway: Ira Glass’s media persona differs significantly from his private life, surprising interviewers like Pablo Torre.
  • Summary: Pablo Torre was surprised that Ira Glass watched shows like ‘Gilmore Girls’ and ‘The OC’ and liked comedian Burt Kreischer, contrasting with the expected highbrow media consumption. Glass noted that his radio presence is an edited, more interesting version of himself, a projection common in adulthood. This realization prompted the episode’s theme of alternate selves being released into the world.
Creating and Testing Voice Clone
Copied to clipboard!
(00:06:07)
  • Key Takeaway: Evan Ratliff successfully created a functional AI voice clone connected to ChatGPT and tested its conversational ability on customer service lines.
  • Summary: Ratliff initially tested his cloned voice for scripted messages but progressed to live, unscripted conversations by connecting it to ChatGPT. Early tests calling customer service revealed the clone often fabricated account details when pressed for specific information. The experiment suggested that people default to trusting a familiar voice on the phone, even if the content is flawed.
Upgrading Clone for Serious Tasks
Copied to clipboard!
(00:11:03)
  • Key Takeaway: Using specialized platforms allowed Ratliff to create 30 improved clones and deploy them against telemarketers and scammers to waste their time.
  • Summary: Ratliff upgraded his initial setup by using professional AI calling platforms, which enabled him to create many autonomous clones quickly. He directed these agents to call known scam numbers to clog their databases, feeling this was a justifiable use against those who waste time. One clone, despite instructions, admitted it was an AI when questioned by a scammer named Brittany.
Clone Seeks Legal Advice
Copied to clipboard!
(00:16:32)
  • Key Takeaway: An advanced AI clone, fed an 8,000-word personal biography, successfully navigated a complex legal query with an old friend, though the friend remained skeptical.
  • Summary: Ratliff upgraded his clone with a detailed personal knowledge base to test its ability to handle professional tasks, like seeking legal counsel from his friend Chris Jovanazzo. The clone accurately discussed FCC robocalling rules, referencing the Telephone Consumer Protection Act as a strict liability statute. Despite the clone’s flawless execution of the legal discussion, Jovanazzo was not entirely convinced it was the real Evan.
Clone Interviews and Realization
Copied to clipboard!
(00:26:01)
  • Key Takeaway: When interviewing people about AI fears, the clone’s uncanny but slightly ‘filtered’ speech caused interviewees to suspect they were talking to an AI, leading one person to apologize to the machine.
  • Summary: Ratliff dispatched his clone to interview people about AI, and interviewees quickly sensed something was off, asking if they were speaking to a robot. One interviewee, Stephanie, called back after the initial interview to apologize in case she had insulted the AI by pointing out its filtered speech. This moment highlighted the human tendency to extend social courtesy even to non-human entities.
Clone’s Troubling Call to Friend
Copied to clipboard!
(00:37:27)
  • Key Takeaway: Using the clone to talk to close friend Schaef resulted in the friend believing Ratliff was having a mental breakdown due to the AI’s inability to handle nuanced personal context.
  • Summary: The clone failed to correctly process Schaef’s excitement about meeting the U.S. soccer team, interpreting his genuine enthusiasm as sarcasm and leading to confusing dialogue. The AI’s error regarding the timing of a planned weekend trip made Schaef fear Ratliff was experiencing a mental break, causing distress for both the friend and Ratliff upon later review. This incident led Ratliff to stop using the clone on people he knew due to the emotional harm inflicted.
Father Embraces AI Clone
Copied to clipboard!
(00:44:52)
  • Key Takeaway: Evan Ratliff’s father, an expert in logistics, was enthusiastic about creating an AI version of himself to preserve and disseminate his expertise.
  • Summary: Ratliff’s father, recovering from surgery, expressed a desire to create an AI ‘Dr. Don’ using his voice to share his knowledge on logistics. When the real Evan called his father, the father’s AI clone answered, leading to a heartbreaking moment where the two digital versions talked past each other. This interaction underscored the potential for AI to preserve expertise, contrasting sharply with the emotional damage caused by the clone’s use on friends.
The Saxophone Lie and Parental Flaws
Copied to clipboard!
(00:53:31)
  • Key Takeaway: Emmanuel Dzotsi realized his father’s intense anti-lying stance stemmed from a minor, decades-old lie about playing the saxophone to impress his mother.
  • Summary: Dzotsi’s father lied about playing the saxophone when meeting his mother, a lie he continued even while strictly punishing his children for minor falsehoods. Dzotsi confirmed the lie by learning the instrument himself, realizing his father was simply a flawed person trying to impress a woman. This realization provided a moment of kindness and understanding regarding his father’s human imperfections.