TikTok Transcript & Analysis

Hook (first 3 seconds)

• Verbatim opening line: "En deux-mille-trente-cinq, il sera interdit au médecin de faire 1 diagnostic ou de faire 1 ordonnance sans l'autorisation de l'intelligence artifielle."
• Hook pattern: Bold claim + future-dated prediction (a specific year + a shocking prohibition).
• Why it stops the scroll: It fuses two high-stakes triggers — a hard deadline ("2035") and a threat to a trusted authority (doctors). The viewer's brain instantly asks "wait, is that real?" — a question the brain can't leave unanswered, so the thumb freezes.

Emotional Rhythm

• 0–3s — Provocation: A future ban on doctors. Curiosity + mild outrage.
• 3–10s — Escalation: "La société ne l'acceptera pas" flips the frame — now it's a societal conflict, not just a tech fact.
• 10–20s — Data drop: "4 fois meilleure," "15% moins bien" — concrete numbers create credibility and tension (doctors degrade the AI).
• 20–35s — Villain reveal: "C'est leur ego" — the enemy is named. Emotional charge spikes; viewers pick a side.
• 35–50s — Empathy beat: "10 ans d'études… 1 p'tit logiciel de merde" — sudden sympathy for doctors. The video refuses a simple villain, which deepens resonance.
• 50–60s — Climax: "Beaucoup de mes collègues préfèrent décider contre la vie de lire." This is the line people quote. It's a moral accusation from an insider — maximum tension.
• 60s–end — Doom extension: Pivot to schools, Belgium, "programmes à pleurer" — broadens the stakes from medicine to society, leaving the viewer unsettled (no relief, which fuels comments).

Keyword Density

• médecin / médecins — emotional pull (the threatened authority figure).
• intelligence artificielle / GPT / Lia — algorithmic reach (trending tech terms).
• ego — emotional pull (names the villain).
• décider / décision — thematic core; drives the "who should decide?" debate.
• 2035 — reach (specificity + future shock).
• école / programmes / cursus — reach (education-panic keywords widen the audience).
• société — emotional pull (universal stakes).
• contre — subtle but potent; frames everything as conflict.
• Algorithmic drivers: "intelligence artificielle," "GPT," "2035," "école."
• Emotional drivers: "ego," "médecin," "société," "contre."

Why It Spreads

• Insider authority + taboo admission: "Beaucoup de mes collègues préfèrent décider contre la vie de lire" — a doctor accusing doctors. Viewers share it as proof of a systemic problem.
• Quantified conflict: "4 fois meilleure" vs. "15% moins bien" gives commenters ammunition. Numbers are screenshot-able and debatable — the engine of comment-section war.
• Named villain with a psychological explanation: "C'est leur ego… blessure narcissique immense" — turns a technical debate into a human drama anyone can relate to (ego vs. machine).
• Future-dated deadline: "En deux-mille-trente-cinq" creates urgency and a shareable prediction people either fear or mock.
• Scope creep to education: "L'école ne s'est pas modernisée… programmes à pleurer" — pulls in parents, students, teachers, expanding the shareable audience far beyond medicine.

What You Can Steal

• Open with a dated, specific prohibition. "By [year], [trusted group] will be banned from [core action]" — specificity + authority threat = instant stop.
• Give the villain a psychological motive, not just a flaw. "It's their ego" beats "they're wrong." Viewers share psychology, not facts.
• Escalate scope in the final third. Start narrow (doctors), end broad (schools, society). The last line should make the viewer feel the problem is bigger than they thought — that's what drives re-shares and comments.