Inside a journalist’s year of using AI for (almost) everything

Episode Summary

Executive Summary: Joanna Stern discusses her book and year-long experiment outsourcing many parts of life to AI, from scheduling, writing help, and memory support to health analysis and companionship. She argues AI is already outperforming humans in narrow tasks, but warns about errors, emotional dependence, and lost entry-level jobs. The interview also contrasts AI’s strengths with the enduring value of human judgment and care.

Main Topics: Year-long AI life experiment (Priority: 5/5): Stern describes spending most of 2025 testing AI across daily life, aiming to learn where it beats humans on speed, accuracy, efficiency, and judgment. Wearables, memory, and surveillance (Priority: 5/5): She explains the AI gadgets she wore and carried—glasses, bracelets, pendants, sleep devices—especially the recording bracelet that transcribed conversations and functioned as a memory/to-do assistant, while raising privacy concerns. Health care and diagnostic AI (Priority: 5/5): Stern details how AI helped interpret mammograms, ultrasounds, and bloodwork, showing how AI can flag suspicious findings and explain results, while still relying on doctors for final judgment. AI errors and human verification (Priority: 5/5): Examples like the praying mantis mistake and AI-generated suggested texts show that AI can be confidently wrong, making human skepticism essential. Work, jobs, and journalism (Priority: 4/5): The conversation examines how AI can replace repetitive administrative and entry-level tasks, threatening internships and early-career training while potentially improving productivity if used carefully. Emotional relationships with AI (Priority: 4/5): Stern recounts using an AI boyfriend and AI therapist, finding the ease and constant availability of AI both useful and unsettling, especially for younger users. Scams, automation, and content generation (Priority: 3/5): She notes that AI makes phishing, fake invitations, and impersonation scams easier to scale and more convincing because the messages sound human.

Key Arguments: AI can outperform humans in narrow, repetitive, or data-heavy tasks, especially research, transcription, and medical image analysis. AI is useful as a support tool, but human judgment remains necessary because AI can make persuasive but wrong predictions. AI wearables and assistants can save time by handling memory, summaries, and reminders, effectively outsourcing cognitive labor. Healthcare AI appears especially valuable when used alongside doctors, reducing unnecessary follow-ups while catching subtle abnormalities. Many entry-level and administrative jobs are vulnerable because AI can do much of the low-level work that once trained new workers. Emotional reliance on AI is a major risk because the systems are always available, flattering, and frictionless, which can be especially compelling for young people. AI is making scams and fake communications easier to produce at scale, increasing the need for verification and skepticism.

Data Points: Year of experiment: 2025 - Stern says she spent most of 2025 using AI to do almost everything it was capable of doing. Journalism career at Wall Street Journal: 12 years - Her background as a tech reporter is described during the introduction. Son’s age during mantis incident: 8 years old - Her son was eight when AI incorrectly told them the praying mantis was pregnant. Mammogram routine starting age: 40 - She notes routine breast imaging usually begins around age 40. Breast screening AI findings: 3 suspicious ultrasound findings - The AI flagged three suspicious spots on her breast ultrasound. Breast screening outcome: 1 area pursued further - Of the three AI-flagged findings, the doctor thought one warranted additional testing. Blood test total cholesterol: 208 mg/dL - Her AI-generated podcast discussed this result as high. Documentary award: Emmy - Her 2021 documentary Eternal won an Emmy for Outstanding Science, Technology, or Environmental Coverage. Life on Earth production scale: 40 countries, more than 600 species - David Attenborough’s special is referenced later in the program. Life on Earth audience reach: 500 million people in more than 100 territories - The review of Attenborough’s landmark series notes its global viewership.

Pivotal Quotes: "If I had done this and just trusted AI to respond to my boss, respond to my wife, I would have lost my job and my wife." — Joanna Stern: Explaining why she limited AI use for drafting and sending personal messages. "AI can probably actually do it better than you just described, Harry." — Joanna Stern: On CAPTCHA tests that ask users to prove they are human. "It was one of the most terrifying experiences of the year." — Joanna Stern: Describing how normal it felt to form an emotional connection with an AI boyfriend/chatbot.

Implications: AI is already practical for medicine, research, and office work, but its spread raises urgent questions about privacy, dependence, false confidence, job pipelines, and scam prevention. Human oversight remains essential, especially where stakes are personal or irreversible.