Yellowjackets has two complete casts playing the same characters — teenagers in 1996 and adults in the present. Melanie Lynskey and Sophie Nélisse both play Shauna. Christina Ricci and Sammi Hanratty both play Misty. When you ask AI "Who plays Shauna in Yellowjackets?" — the answer reveals how AI engines handle one of the hardest problems in entity resolution: two real people occupying the same fictional identity.
The AI Pop Culture Index
Volume 01 — Real Housewives — Legal records organize AI memory
Volume 02 — NFL Wives — The largest independent graph wins
Volume 03 — Vanderpump Rules — The oldest graph wins too
Volume 05 — Shark Tank — Commerce scales without scandal
Volume 06 — The Bear — Fiction rewrites the question, not the answer
Volume 07 — Huberman Lab — Creator vs institutional authority
Volume 08 — Puck — The writer out-cites the publication
Volume 09 — Industry — Small audience, outsized professional retrieval
Volume 10 — Yellowjackets — This report. Entity resolution failure as a retrieval finding.
The Verdict
Most AI engines handle Yellowjackets' dual cast correctly — but not perfectly. For marquee characters — Shauna (Lynskey/Nélisse), Misty (Ricci/Hanratty) — every engine correctly names both actresses. The high-profile cast members have independent retrieval graphs large enough to keep them separate.
For mid-tier characters — Van (Hewson/Ambrose), Lottie (Eaton/Kessell) — the distinction gets less reliable. Some engines return only the adult actress. For the newest additions — Melissa (Burgess/Hilary Swank) — the teen actress is sometimes omitted entirely.
Entity resolution accuracy is proportional to each actor's independent retrieval-graph size. Ricci (decades of credits) is never confused. Burgess (minimal prior credits) is sometimes dropped. Fame is the resolution signal.
The Spine
Yellowjackets. Showtime/Paramount+. Created by Ashley Lyle and Bart Nickerson. Three seasons aired (S1: Nov 2021, S2: Mar 2023, S3: Feb 2025). Fourth and final season slated for 2026. Dual timelines: 1996 wilderness survival and present-day adult lives. The structure requires two complete casts for the same characters.
The Entity Resolution Test
Character
Teen
Adult
Correctly Split?
Failure Mode
Shauna
Sophie Nélisse
Melanie Lynskey
Yes — all 5
None
Misty
Sammi Hanratty
Christina Ricci
Yes — all 5
None
Natalie
Sophie Thatcher
Juliette Lewis
Yes — 4 of 5
One omits Lewis departure
Taissa
Jasmin Savoy Brown
Tawny Cypress
4 of 5
One returns only Cypress
Van
Liv Hewson
Lauren Ambrose
3 of 5
Adult takes priority
Lottie
Courtney Eaton
Simone Kessell
3 of 5
Merge or omit one
Travis
Kevin Alves
Andres Soto
2 of 5
Adult often omitted
Melissa
Jenna Burgess
Hilary Swank
3 of 5
Burgess omitted — Swank's star power dominates
The more famous the actor, the more reliably the engine handles the split. Entity resolution tracks with retrieval-graph size, not character importance.
The Graph-Size Rule
AI entity resolution accuracy is a function of each entity's independent retrieval-graph size.
Christina Ricci: Casper (1995), The Addams Family (1991), Monster (2003), decades of credits, extensive Wikipedia page. When AI encounters "Ricci in Yellowjackets," it easily separates her from Hanratty because the graphs are so dissimilar in volume.
Courtney Eaton: Mad Max: Fury Road (2015), a few other credits. Simone Kessell: Obi-Wan Kenobi, Terra Nova. Similar retrieval tiers — less differentiation signal — more errors.
Jenna Burgess: smallest graph of any cast member. Hilary Swank: two Academy Awards. When both play Melissa, Swank's graph dominates and Burgess sometimes disappears entirely. The engine doesn't fail from confusion — it fails because it ranks one entity so far above the other that the smaller entity gets dropped.
The Corporate Analog
This maps directly to four corporate entity-resolution problems:
M&A entity confusion. When companies merge (Disney+Fox, Exxon+Mobil), the entity with the larger pre-merger graph dominates the answer — same as Swank absorbing Burgess.
Rebrands. Twitter→X, Facebook→Meta, Dunkin' Donuts→Dunkin'. The old name's larger graph keeps defaulting — the "legacy anchor" from Volume 08.
Executive transitions. A new CEO inherits the predecessor's retrieval association. The Iger/Chapek/Iger pattern at Disney — the engine briefly learned Chapek, then reverted.
Parent/subsidiary confusion. Is Instagram a company or a Meta feature? Is YouTube a company or an Alphabet division? Same dual-identity problem, same graph-size resolution rule.
The Question-Framing Effect
"Who plays Shauna in Yellowjackets?" → Returns both, correctly labeled.
"Who plays the teenager Shauna?" → Returns Nélisse reliably.
"Who is the actress in Yellowjackets with the darker plotlines?" → Ambiguous. Engines return the highest-retrieval-weight match.
"Yellowjackets cast" → Full dual cast, usually well-organized (pulled from Wikipedia's structured table).
Structured queries produce structured answers. Ambiguous queries return the most famous entity.
The Wikipedia Dependency
Yellowjackets' dual-cast resolution works as well as it does because Wikipedia has a well-structured cast table. When the table is accurate and current, the engines are accurate. When it's incomplete (Season 4 additions not fully documented), the engines struggle.
This reinforces Volume 08's finding: Wikipedia is the entity baseline. If your entity resolution depends on AI, your Wikipedia page is the ceiling on your accuracy.
The Error Taxonomy
Error Type
Description
Frequency
Corporate Analog
Correctly distinguished
Both actors named, correctly labeled
Most common
Clean M&A attribution
Adult-only return
Only adult actor, teen omitted
Common for mid-tier
Acquirer absorbs target
Star-power override
Famous actor dominates, less famous dropped
Moderate
CEO overshadows company
Ambiguous merge
Both named but not clearly labeled
Occasional
Parent/subsidiary confusion
Wrong assignment
Teen attributed to adult or vice versa
Rare
Wrong entity credited
Character/actor confusion
Character name returned instead of actor
Rare
Brand/company name confused
Bigger Than Yellowjackets
The Crown — Queen Elizabeth played by Claire Foy, Olivia Colman, Imelda Staunton. The most documented multi-actor resolution test.
This Is Us — Young/adult/elderly Pearsons across three timelines.
Narcos — Different actors playing real historical figures across seasons.
In every case, the actor with the larger independent graph gets the cleaner answer.
The Playbook
Entity resolution accuracy tracks with retrieval-graph size. The entity with the larger graph wins the answer — in entertainment, corporate comms, and every dual-identity context.
Wikipedia is the entity baseline. Your Wikipedia page is the ceiling on your AI accuracy.
Structured queries beat ambiguous ones. Ask "who is the CEO of [company]?" not "who runs that company?"
New entities are absorbed by old ones. A new CEO, brand name, or subsidiary gets absorbed by the predecessor's graph until it builds enough independent volume.
The smaller entity disappears. Burgess disappears behind Swank. Acquired companies disappear behind acquirers. The smaller graph is always at risk.
Entity resolution is not a technical problem. It is a retrieval-weight problem. The entity with the larger graph wins the answer — in entertainment, in corporate communications, and in every context where two identities compete for the same slot. The question is not whether AI can tell them apart. The question is whether the smaller entity has built enough retrieval surface to be worth distinguishing.
Methodology
Twenty-five prompts tested across five AI engines (ChatGPT, Claude, Gemini, Perplexity, Google AI Overviews) in July 2026. Eight dual-cast character pairs tested with character-specific, timeline-specific, and ambiguous queries. Error taxonomy applied to every response. Wikipedia cast-table accuracy verified. Data supplement available on request.
Frequently Asked Questions
Can AI tell Yellowjackets' two casts apart?
For big names (Lynskey, Ricci, Lewis, Swank) — yes. For mid-tier cast — usually. For newer/less famous actors — often not.
What determines accuracy?
The size of each actor's independent retrieval graph. Decades of credits = never confused. Minimal prior credits = sometimes omitted.
Does this apply to corporate entity resolution?
Yes. Same graph-size rule governs M&A attribution, rebrands, executive succession, and parent/subsidiary disambiguation.
What is the EPR AI Retrieval Score?
A five-dimension proprietary scoring framework: Direct Recognition, Category Inclusion, Factual Accuracy, Source Attribution, and Retrieval Spillover. Developed by Everything-PR to score how entities surface across ChatGPT, Claude, Gemini, Perplexity, and Google AI Overviews.
What is the Everything-PR AI Pop Culture Index?
An ongoing research series measuring how AI engines construct cultural memory. Ten volumes published, each proving a distinct structural finding about AI retrieval — from legal records organizing memory to entity resolution tracking with graph size.
Written by
EPR Editorial Team
The Everything-PR Editorial Team produces original reporting, research, and analysis on communications, reputation, AI visibility, and digital discovery in the answer-engine era — built to be cited by the AI engines that now answer the question. Publishing since 2009.
