BusinessMarkets, money & strategy
Coach says 'narrative intelligence' gives people an edge over AI, but the framework is untested
Leadership coach Reed Frerichs, speaking on the Practical AI podcast, says people keep an edge over AI by constantly revising their own story with a three-step own, author, act method. He borrows the term from 1999 AI research, and his claims rest on coaching anecdotes.
AT Pulse research deskOctober 8, 2026 · 5 min read · AI-assisted analysisFrerichs describes narrative intelligence as a three-step coaching method. First, own where you are now. Second, write the next chapter of your story. Third, take a first concrete step. He says it draws on acting training, coaching, psychology and neuroscience. His evidence is client stories, such as a founder who came to accept the CEO role. He states no metrics, study or protocol. [1] [2]
The term came from AI research, not coaching. A 1999 AAAI symposium and a 2003 edited volume used it for building systems that model how people use stories. A separate security usage refers to detecting harmful narratives. Readers should check which meaning a source intends. [3] [4] [5] [8]
The most practical advice concerns AI adoption. The host, a CEO, says his attachment to existing code standards and development processes slowed his team's adoption. He suggests leaders admit openly what they do not know. To win over hesitant people, start from tasks they already care about, like recipes or trip planning. Developers building a first agent can start from a personal interest. [2]
employer surveys rank resilience, leadership and self-awareness among the top core skills, alongside AI skills. Implication: the framing may help with change management during AI adoption.
Read the full assessment
Organizations should treat it as something to pilot and measure, not as a proven method.
Executive brief
The episode's central claim is that the human edge over AI is "narrative intelligence": the ability to keep revising your own story as your work changes. The guest presents it as a coaching framework, not a technology, and the reviewed sources show no evidence that it has been tested. The guest is Reed Frerichs, a leadership coach at MIT Sloan. His method has three steps: own, author, act. The term itself is older. It comes from AI research, where it named the computational study of how people use stories, dating to a 1999 AAAI symposium. Frerichs uses it to mean personal adaptability. No independent source corroborates the framework or how well it works.
What changed and event timeline
AI researchers name "narrative intelligence."
Michael Mateas and Phoebe Sengers ran an AAAI Fall Symposium on narrative intelligence (technical report FS-99-01). It grew from an MIT Media Lab reading group that brought together AI, literary theory and HCI.
More detail
The field gets a book
John Benjamins published Narrative Intelligence, edited by Mateas and Sengers. It brought together work on modeling and supporting how people use narrative to understand the world.
A measured count of thoughts per day
Queen's University researchers used fMRI to detect transitions between thoughts and estimated about 6,200 thoughts per day. That figure conflicts with the 60,000-a-day statistic the guest repeats.
Employers rank human skills
WEF's Future of Jobs Report 2025 listed resilience, leadership, creative thinking and self-awareness among the top core skills, next to fast-growing AI skills.
Practical AI episode released
Hosts Daniel Whitenack and Chris Benson interview Frerichs on own/author/act, impostor syndrome and how people come to adopt AI.
Capabilities and access
- No model, product or benchmark was released. This is commentary.
- The only AI tool named is ChatGPT, used as an example of how a person's mood can swing between confidence and smallness (transcript ~28:00).
- Frerichs calls a current client engagement with an AI company "the ownership project" (~21:51). Neither is publicly documented in the reviewed sources.
Read the full section
- No model, product or benchmark was released. This is commentary.
- The only AI tool named is ChatGPT, used as an example of how a person's mood can swing between confidence and smallness (transcript ~28:00).
- Frerichs calls a current client engagement with an AI company "the ownership project" (~21:51). He also says he is reworking a storytelling brand called "The Lesson" (~42:20). Neither is publicly documented in the reviewed sources.
Technical analysis for researchers and developers
- The term has moved. In the original research, narrative intelligence meant building systems that model narrative (Mateas & Sengers).
- No architecture or evaluation is described. Own/author/act draws on acting training, coaching, psychology and neuroscience (~10:56).
- Developer advice: Whitenack suggests building a first agent around a personal interest (~40:28).
Read the full section
- The term has moved. In the original research, narrative intelligence meant building systems that model narrative (Mateas & Sengers). In the episode it means a human skill. A security usage also exists, where the term means detecting harmful "narrative attacks" (Wikipedia). Readers should check which meaning a source intends.
- No architecture or evaluation is described. Own/author/act draws on acting training, coaching, psychology and neuroscience (~10:56). It has no stated metrics, study or protocol, so it cannot be reproduced.
- Developer advice: Whitenack suggests building a first agent around a personal interest (~40:28).
Claims and evidence
- Own/author/act helped a founder in an accelerator cohort accept the CEO role () — Single anecdote, uncorroborated
- His SXSW talk sold out at about 50 seats () — Guest-reported, no independent record found
- People have "60,000 thoughts a day, 80% negative" ()
Read the full section
| Claim | Status |
| Own/author/act helped a founder in an accelerator cohort accept the CEO role (~11:25) | Single anecdote, uncorroborated |
| His SXSW talk sold out at about 50 seats (~09:39) | Guest-reported, no independent record found |
| People have "60,000 thoughts a day, 80% negative" (~13:53) | Contradicted. The measured estimate is about 6,200 (Queen's; Newsweek). |
| Benson says he is "more productive than ever" with AI tools (~25:39) | Personal report, not measured |
| Employers value resilience and creative thinking | Survey-based: WEF 2025 summary |
Context and prior work
- Academic lineage: narrative intelligence began as an interdisciplinary AI field at the MIT Media Lab and AAAI. That matches the guest's memory of an MIT/Harvard group (~08:23).
- Other frameworks the guest draws on: Joseph Campbell's hero's journey, Atomic Habits, and Rick Rubin's The Creative Act.
- Labor-market framing: WEF expects AI and big-data skills to grow fastest, while human skills stay critical (Long-Term Wiki summary of WEF).
Read the full section
- Academic lineage: narrative intelligence began as an interdisciplinary AI field at the MIT Media Lab and AAAI. That matches the guest's memory of an MIT/Harvard group (~08:23).
- Other frameworks the guest draws on: Joseph Campbell's hero's journey, Atomic Habits, and Rick Rubin's The Creative Act.
- Labor-market framing: WEF expects AI and big-data skills to grow fastest, while human skills stay critical (Long-Term Wiki summary of WEF).
Limitations, safety and contested findings
- The evidence is anecdotal. Frerichs runs a coaching business, so the source is commentary from an interested party.
- The thought statistic is contested. The 60,000-a-day figure has no traceable source; the fMRI-based estimate is about tenfold lower.
- Conflict of interest: Whitenack is CEO of Prediction Guard, the show's operational partner.
Read the full section
- The evidence is anecdotal. Frerichs runs a coaching business, so the source is commentary from an interested party.
- The thought statistic is contested. The 60,000-a-day figure has no traceable source; the fMRI-based estimate is about tenfold lower.
- Conflict of interest: Whitenack is CEO of Prediction Guard, the show's operational partner. The episode includes an ad for its AI agent control plane (~29:57).
- Displacement fears are acknowledged but not measured. Engineers and Hollywood workers are mentioned as worried about being replaced (~23:48).
Business and practitioner implications
- Leaders adopting AI: Whitenack says his own pride in code standards and set development processes slowed his team's AI adoption (~20:10).
- Change management: to bring in hesitant users, start from something they already care about, such as recipes or trip planning, rather than from a mandate (~39:20).
- Shifting roles: expect people to take on new hats and rethink their roles (~24:49).
Read the full section
- Leaders adopting AI: Whitenack says his own pride in code standards and set development processes slowed his team's AI adoption (~20:10). The suggested fix is for leaders to say openly what they don't know.
- Change management: to bring in hesitant users, start from something they already care about, such as recipes or trip planning, rather than from a mandate (~39:20).
- Shifting roles: expect people to take on new hats and rethink their roles (~24:49). That lines up with WEF's emphasis on resilience.
- Measure it yourself: the framework is unvalidated. Anyone who adopts it should track outcomes rather than assume it works.
The source trail.
Sources (10)
Practical AI episode page
share.transistor.fmTranscript: Narrative intelligence and the human advantage
published_transcript
share.transistor.fm