Story and writing

Story structure, screenplays, series bibles, characters, dialogue, narrative evaluation

• articleWhich open-source screenwriting apps (Trelby, Starc, Beat, afterwriting) read and write FDX and Fountain, and what licence does each ship under?
  Trelby 2.4.16.2 licence statements disagree: LICENSE is GPL v2 text, setup.cfg declares GPL-3.0-or-later and debian/copyright says GPL-2 (curator-read pinned bytes); afterwriting v1.17.3 README states MIT. The Trelby conflict is preserved, not resolved.
• articleWhat do the ink JSON runtime format (v1.2.1) and the Twine 2 HTML/archive format preserve about line identity, localisation and variables?
  At ink 35c63e52 the JSON-format document example shows inkVersion 10 while the runtime defines inkVersionCurrent = 21; the JSON format document defines no localisation or line-identifier field in the sections read.
• articleWhich failure modes does Chapter XIII (1919) name for scripting what cannot be photographed or staged, and which are the authors' claims versus examples versus quoted voices?
  Five script-side failures are named: a scene that will not register; one that is possible but impractical (Sargent's chase with dust and smoke); a demand for a particular background; a demand for an effect whose cost is out of proportion; and a belief that the camera does everything it seems to (trick photography versus careful work). The staging failure of players leaving the frame, and the camera and tripod hazards (Fildew), describe studio and camera problems, not script mistakes. The remedy the authors give is to write difficult scenes as a suggestion. All of this is 1919 trade advice; the voices are separated below. The chapter says nothing about modern generative failure modes.
• articleWhat check can an agent run on a draft or outline against an intended audience before production, built from Baker (1919, Chapter X) and labelled 1919 stage practice and a proposal?
  PROPOSAL built on 1919 stage practice. Baker gives no checklist; this batch assembles one from Chapter X (labelled synthesis), in this order: (0) freeze the central human and artistic purposes; (1) state the intended audience, its knowledge and approach, and whether it is general or special; (2) interest check; (3) comic/tragic and sympathy check; (4) motivation check for informed viewers; (5) form and length check; (6) order check; (7) truckling triage: any audience-driven change that alters the purposes is refused, and an audience-unfriendly element is cut only after a recorded attempt to make the audience feel it; (8) record each change with reason and stage; (9) readiness gate. It is run as a separate critical pass, in line with Baker's advice to keep creative and critical work apart. Baker supplies no outside-audience protocol; any LLM or human reader panel is an untested addition.
• articleWhat workflow can an agent follow to outline a story general-outline-first and then fill in episodes, with each step labelled as classical theory (in translation) or as proposal?
  A nine-step workflow, each step tagged. Classical theory (in translation, Butcher XVI-XVII): outline the general chain first and mark what stays outside the plan (1); leave the mode of any recognition open (2); then give names (3); then fill in episodes and test each for relevance to the action (4); rank each reveal by origin (5); picture each scene to catch inconsistencies (6); act out emotional speech (7). Proposal (this batch): record the outline and episodes as separate, linked layers with a change log, since later work may revise the outline (8, labelled proposal; Baker's flexibility is in C3-B194-S47-02 by ID only), and run a mapping check that every episode points to an outline beat (9). No source documents an LLM doing any of this.
• articleGiven a scene description, what check can an agent run so every line shows something that can be seen (1914 practice), and what in that check is only the lead's proposal, including any use for text-to-video prompts?
  Agent check assembled from Phillips (1914 practice, each step tagged with its source) plus lead-proposed steps (tagged PROPOSAL). Steps: strike 'registers <emotion>' lines; replace each with a restrained visible sign; add a short beat only where the feeling needs room; use caption/close-view/vision-scene for dreams (period devices); write sounds only when they bear on the action, as source plus reaction; give each line an action verb and only visible objects (outside Chapter V); finish with a reader test (editor/director/actor in 1914; a second reader or agent listing what they would see in the PROPOSAL). Use for text-to-video prompt writing is a hypothesis only; no source inspected documents it, and Phillips does not address AI.
• articleWhat is a concrete checklist for revising a dialogue scene from Baker's 1919 chapter, and how could it be adapted, as an untested proposal, to screen scripts and LLM-assisted revision?
  Nine revision steps follow directly from Baker's Chapter VIII (fact check, job pass, swap test, author-voice check, cut detail, split blocks, delivery and speakability, pantomime check, naturalness read). They are labelled 1919 stage practice. A second list of five proposals adapts them to screen and LLM-assisted use and is untested: Baker's stage aids are supplied by actors, so screen adaptation must say what image and performance carry; an LLM could run the swap test as a blind attribution pass and propose cuts, but no source in this batch tests that. Existing rows on voice distinctiveness and LLM judging (C3-B119-S26, C3-B217-S52) give measured limits that a user of the proposal should read first.
• articleWhat is documented about platform AI policy toward authors in Chinese web fiction, and what does platform economics say about serial incentives?
  Thin, secondary. One news item (AsiaIPLaw, 26 July 2024) reports that ByteDance's Fanqie Novel added an AI-training supplement to author contracts (works usable for training, platform copyright over generated content, priority beta access) and offered a termination feature on 23 July 2024; it also states the platform's revenue share rewards number of readers and reading time. The Rest of World 2026 article on platform restrictions of AI-generated fiction could not be fetched (HTTP 429), and Qidian/Webnovel AI rules were not found; later 2026 developments appear only in search summaries and are not evidence here. No source read gives chapter length, update cadence or hook rules.
• articleHow does CoSER get LLMs to write multi-character dialogue and judge it, and how could an agent add the missing speaker-blind human or model check for character voice?
  CoSER's given-circumstance acting (GCA) protocol has one LLM instance per character write speech, action and private thought, an NSP model picks speakers, an environment model replies, and a penalty-based critic lists flaws against the original conversation. Its Character Fidelity rubric covers language style but is judged by an LLM and not separately checked against humans in the text read. A speaker-blind attribution test on GCA outputs is not part of the paper; it is offered below as an untested proposal.
• articleWas CoSER's automatic LLM judge validated against its human raters, with what statistic and numbers, and how far can an agent trust that validation for judging character dialogue?
  Yes, partly. Table 5 reports 'alignment' = percentage agreement with the human raters on which of two models is better, ties dropped. GPT-4o as judge gives 68.6, DeepSeek-V3 65.1 and DeepSeek-R1 77.5; BLEU gets 75.3 and ROUGE-L 72.0, so only the R1 judge beats the n-gram metrics. No correlation coefficient, confidence interval or inter-annotator agreement is given. The paper's main results use the lower-aligned GPT-4o judge. The check is at the whole-score level with 7 models, not per rubric dimension such as Character Fidelity.
• articleDid people rate the AI-written character dialogue in CoSER, what exactly did they score, and does that overturn the earlier finding that no human-rater study of LLM-generated lines attributed to characters exists?
  Yes, with limits. CoSER v3 Section 5.3 reports three annotators who each scored 20 of 60 CoSER Test samples, rating the simulated conversations of 7 LLMs on a 1 to 10 scale while seeing the background context and the original conversations. The text read names no criterion beyond 'GCA simulation' quality (role-playing ability overall); it is not stated as character fidelity, distinguishability or voice, and no who-said-it task appears in the text read. CoSER-70B scored highest (6.783, win rate 86.9%). The earlier scoped negative is confirmed for who-said-it attribution and voice distinctiveness, and narrowed (not overturned) for human rating of LLM-generated character dialogue in general.
• articleWhich documented techniques keep an LLM from giving a character information that character should not have, and what effect did each show?
  Six techniques have measured effects, each in a different setting, so they are not comparable head-to-head. (1) Two-pass perspective filtering (SimToM): GPT-3.5-Turbo ToMi false-belief 81.0 (+13.8 over 0-shot); single-prompt version loses the gain; human-annotated perspectives reach 96/96/82/98; mixed on OpenToM (candidate 04). (2) Per-character belief graphs (SymbolicToM): GPT-3-Davinci on self-built structure sets 20/25/39 to 92/91/90; needs chronological stories and suffers from extraction errors. (3) Visibility-tagged fact memory with scene anchoring (ReverieMem): refusal 81.2 versus 8.7 to 11.3 for retrieve-everything RAG; ablations show both layers matter; one speaker only. (4) Separate agent contexts versus an omniscient script (Zhou et al.): fewer early leaks (0.39 vs 0.13) but far fewer deals (30 vs 94 percent). (5) Awareness reminder (SimpleToM): behaviour prediction above 80 for three models, judgment low. (6) Key Facts Injection: 73 percent fewer 'known-information forge...
• articleHow have unearned character knowledge or leaked information been detected and rated in LLM-generated stories and simulated conversations?
  Two partial measurements exist in the sources inspected; no source measured character-to-character secret leakage in generated fiction with human raters (bounded negative). (1) ConStory-Checker (arXiv 2603.05890v1) is a four-stage LLM judge with a 'Knowledge Contradictions' sub-type (characters showing knowledge or skills beyond their background, including information they could not have accessed given their background), with quoted evidence and positions; validated on 1,000 planted errors: precision 0.884, recall 0.550, F1 0.678 against human experts at F1 0.281. No sub-type rate is reported in the text read. (2) Zhou et al. (arXiv 2403.05020v4) use the early position of a name in a find-the-mutual-friend dialogue as a leakage proxy: 0.13 for an omniscient single-LLM script versus 0.39 with separate agents (450 simulations per model and mode).
• articleIf an LLM correctly says a character does not know a secret, will it then predict that the character acts as if they do not know?
  SimpleToM (arXiv 2410.13648v2) uses two-sentence everyday stories (a hidden fact; a person acting without knowing it) and asks three questions: is X aware (mental state), what will X do (behaviour), was X's action reasonable (judgment). Models were near ceiling on awareness (GPT-4o 95.6, Claude-3.5-Sonnet 97.9) but far lower on behaviour (49.5, 67.0) and judgment (15.3, 24.9) where chance is 50; o1-preview reached 84.1 on behaviour. The failure in the authors' example is predicting the observer acts on the hidden fact. Reminders of the awareness answer lifted behaviour above 80 for three models but judgment stayed low.
• articleHow deep can nested beliefs ('A thinks B thinks C knows') go before LLM answers fail, and what does deception do to accuracy?
  Hi-ToM (arXiv 2310.16755v1, 2023) generates stories with five agents, rooms, containers and an object, plus optional public or private (deceptive) statements, and asks five questions per story at orders 0 to 4. Four 2023 LLMs scored below 60 percent overall; joint accuracy (correct only if all lower orders are also right) fell from near perfect at order 0 to near zero at order 4; for GPT-4, going from 0 to 4 deceptive sentences per story cost 11 to 32 points of joint accuracy (orders 4 to 1), and chain-of-thought gave no substantial gain. SymbolicToM's authors report a third-order test (their own small set, D2) where per-character belief graphs helped. The Hi-ToM total story count is not stated in the text read.
• articleWhen an LLM reads a conversation where a character was absent for part of it, how often does it credit that character with the missed information, and how consistent are its answers across ways of asking?
  FANToM (Kim et al., arXiv 2310.15421v3) builds 256 multi-party conversations in which a character leaves and rejoins, then asks the same underlying question six ways (free-form belief, multiple-choice belief, answerability list and yes/no, information-access list and yes/no). The model reads everything, so an answer containing what the absent character missed is the failure. In the 2023 models tested, success on one format did not carry to the others: the best chain-of-thought GPT-4 run scored 26.6 on 'all question types' against 87.5 for humans, and the dominant list error was including characters who were unaware. Results are for 2023 models.
• articleWhich story-based theory-of-mind benchmarks test which character knows or believes what, what does each actually ask, and how large is each?
  Five public benchmarks are inspected. ToMi: Sally-Anne object-location stories (999 narratives per OpenToM's table; paper body not read). Hi-ToM: generated stories with five agents and one question each at orders 0 to 4, optional public or private deceptive statements. OpenToM: 696 narratives with 23 first- and second-order questions each (location, multi-hop container/accessibility, attitude). FANToM: 256 conversations in which a character leaves and rejoins; belief, answerability and information-access questions. SimpleToM: two-sentence everyday stories, 1147 per category, with awareness, behaviour and judgment questions. All are template-built or LLM-generated; none uses published fiction.
• articleI want evidence that an AI-written character voice is convincing: what do these three papers actually test with humans, and what is the simplest check I can do first?
  Of the three, only ECHO has people judging AI-generated text, and it asks a different question: which of two answers came from the real (living, ordinary) person, not which character said a line. Deezer's paper has no human raters and no generated text. CoSER may have a human scoring study, but that is unverified here. So a newcomer should not expect any of them to say that readers can tell characters apart in AI-written dialogue. The simplest first check in ECHO's style is a blind pair test on your own cast: one human-written and one AI-written line for the same situation, pick the human one, with 50% as chance; for who-said-it checks use the protocol in S52-08.
• articleDoes the Deezer Research paper "Distinguishing Fictional Voices" include any human evaluation of LLM-generated dialogue, and what does it actually measure about character voice?
  No. The paper (arXiv 2401.16968v1, 30 Jan 2024, Deezer Research and Loria) is fully automatic and uses only human-written novel text. It encodes each character's quotes with pretrained authorship-verification models and tests whether characters can be told apart (character vs character) and whether unseen quotes can be attributed to characters. Chapterwise: LUAR reached AUC 81.6 for character-vs-character but 53.6 for quote attribution, only slightly above chance. The authors guess the attribution task is hard "even for humans" but ran no human test. S52's negative is confirmed for this paper.
• articleDid the ECHO paper have people rate LLM-generated role-play answers against a real person, how many raters and items, what were the results, and were its LLM judges checked against the humans?
  ECHO (arXiv 2404.13957v1, 22 Apr 2024) does include a human evaluation of LLM-generated text: acquaintances of ten real volunteers (at least seven each) saw ten pairs of answers, one from the real person and one from a role-playing LLM, and picked the human one. The headline metric is how often the LLM answer was chosen (50% is chance). Overall rates ran 37.5 to 48.2 across methods, with GPTs closest to chance. The targets are ordinary living people, not fictional characters, and the task is human-versus-machine, not which character said a line, so S52's negative narrows rather than falls. No agreement statistic was found, and LLM judges were scored against ground truth, not validated against the human raters; they picked the LLM answer as the human one above 90% for the two GPT-4 judges.
• articleDo LLM-played characters keep their voice over a long dialogue, how is drift measured, and what do the authors say it converges to?
  Three open-access papers measure drift without human raters in the main loop. Li et al. (arXiv 2402.10962v4) report instruction drift within eight rounds of two-chatbot self-chat using code-based probes. Luz de Araujo et al. (arXiv 2512.12775v2, EACL 2026 per abstract) run over 100-round dialogues with eight RoleBench characters across seven models and report that fidelity (knowledge, style, in-character consistency) falls and persona behaviour reverts toward a no-persona baseline. Abdulhai et al. (arXiv 2511.00222v1) show that a line-to-line consistency metric can rise while prompt consistency falls. The judge validations are small or single-rater.
• articleHow was PersonaGym's PersonaScore validated against human judgement, and what does its Linguistic Habits task say about voice?
  PersonaGym (arXiv 2407.18416v5, 5 Sep 2025) scores persona agents on five tasks with an ensemble of two LLM judges using 1-5 rubrics. A human study of five experts rated 1,500 responses (100 personas, three models) and reports Fleiss kappa 0.71 among annotators and average Spearman 75.1% and Kendall 62.73% between PersonaScore and humans. Linguistic Habits was the hardest task. Personas are described people (for example professions), not fictional characters, and the text and the correlation table disagree on one Kendall claim.
• articleCan an LLM judge reliably say which character is speaking, and what does that imply for using an LLM to check character voice?
  PersonaEval (arXiv 2508.10014v1, 6 Aug 2025) gives LLM evaluators dialogue context plus four or five described candidate roles and asks who the second speaker is. The paper reports the best LLM at around 69% top-1 accuracy against 90.8% for 20 human volunteers on a hard sample, and says models lean on speaking style while humans use addressing and intent. Items are human-authored (English novels, Chinese scripts, English explainer videos), so the result says how judges treat existing text, not how they treat generated dialogue.
• articleHow well can people actually tell which character said a line, and what does that mean for checking my own dialogue?
  Two human studies on human-authored dialogue are in the sources. In Speaker Verification, ten non-expert annotators judging whether two sets of utterances came from the same speaker scored 56.50% accuracy on the base level when shown only utterances (67.48% when shown conversations); in PersonaEval, 20 educated volunteers scored 90.8% when choosing among named candidate roles with profiles and context, on cases chosen because a strong model got them wrong. The gap suggests context cues (who addresses whom, plot) carry most of the human signal. No human study of attributing LLM-generated lines to characters was found.
• articleWhich English-language role-play or persona benchmarks exist, and what does each one actually score (is character-to-character voice distinctiveness among the scored things)?
  Four benchmark-style resources were read. RoleBench scores reference overlap (ROUGE-L) and judge/human win rates for single-turn answers in a role; PersonaGym scores five rubric tasks per demographic persona with an LLM-judge ensemble; PersonaEval scores whether a judge can name the speaker in human-written dialogue; the Speaker Verification work is the only one read that defines a cross-character measure (Distinction Score) on generated lines. CharacterEval is Chinese-only on its abstract page. None of the four reports a human-rated test of whether generated characters sound different from each other.
• articleWhat does open academic research measure about how people watch short-form vertical video, and does any of it test hooks, pacing or framing?
  Three open papers measure short-form viewing without manipulating hooks, pacing or framing: a TikTok data-donation study (arXiv 2301.04945v2, 347 users) reports a median per-view 'attention' (inferred watch fraction, capped at 100%) of 82%; a lab study with a YouTube-Shorts-like player (arXiv 2603.22663v1, 50 users) reports about 70% of sessions ending before 20% of the video; a third (arXiv 2503.20030v1) studies how recommendations evolve. The 82% and 70% figures are different statistics from different populations. The three papers carry different reuse statements (CC BY 4.0, an ACM permission notice, and none), so quote them minimally.
• articleDoes the WGA Screen Credits Manual mention AI-generated material, and how does the MBA's AI clause relate to source-material credit?
  The WGA Screen Credits Manual (effective for notices after 31 December 2021) contains no text on AI. The 2023 MBA Memorandum of Agreement, Article 72.B, says material produced by traditional AI or generative AI (GAI) is not literary material. Article 72.C.3 says that when a Company furnishes a writer with unpublished, unexploited GAI-produced material and instructs its use as a basis, that material is not source material for determining writing credit; the closing paragraphs of 72.C extend this to a writer who uses GAI with the Company's consent while preparing literary material, and such writer-prepared material is considered literary material rather than material produced by GAI. The copy read is the MOA, not the integrated MBA text. PDF text only; not legal advice.
• articleWhat does a more detailed hierarchical outline plus a controller (DOC) add over Re3, how was it evaluated, and how did humans interact with it?
  DOC (Yang et al., v3 June 2023) expands a brief outline into a hierarchical depth-3 outline (2-5 children per node; settings and characters attached; relevance/coherence filtering) and uses an OPT-350m FUDGE-style controller to keep passages faithful to the outline; the drafting model is OPT-175B. In pairwise Surge AI ratings on passages from 20 stories, DOC beat Re3 on coherence (67.6 vs 45.1), relevance (65.3 vs 37.1) and interest (60.1 vs 39.4); ablations credit the outline for interest and both parts for relevance; in a 20-run human-interactive test authors strongly preferred DOC. Leaf-level outline faithfulness remained 58.5% with the controller and 37.8% without.
• articleWhat does ConStory-Bench measure about consistency errors in long LLM-written stories, how does its checker work, and how far can its checker be trusted compared with human readers?
  ConStory-Bench (arXiv v1, March 2026) gives 2,000 prompts for 8,000-10,000-word stories, a 5-category / 19-subtype taxonomy and an o4-mini based four-stage checker that outputs quoted contradictory spans with character offsets. The authors report that errors cluster in factual and timeline dimensions and near 40-60% of the narrative, and that the checker found 55% of 1,000 injected errors against 17.1% for two professional novel writers. That human comparison used LLM-injected errors and two annotators paid $1.00 per story, so it supports checker-as-triage, not checker-as-ground-truth.
• articleHow was FlawedFictions built and annotated, how well did models and humans do at finding injected continuity errors, and what failure modes did the authors report?
  FlawedFictions (arXiv v3, Dec 2025) injects a continuity error into a human-written short story by negating a first-act proposition in later acts, filters with an LLM judge and majority human verification, and scores binary detection plus two-way sentence localization. On 414 stories (average 731 words) the best model rows reach human level (accuracy 0.76) but most models are near chance; on 200 longer stories (average 2,703 words) the best CEEval-Full is 0.53, barely above the always-no-error baseline (0.51). Reported failure modes: over-reading motivations and genre conventions, miscounting entity states, and no gain from more reasoning tokens.
• articleWhat is the architecture of the "Generative Agents" LLM characters (memory, reflection, planning), stated plainly?
  Park et al. (arXiv 2304.03442, pinned v2, UIST 2023) describe an LLM wrapper with a memory stream of natural-language observations, a retrieval function that scores memories by recency, importance and relevance, periodic reflection that writes higher-level memories, and planning that turns the agent's summary and memories into daily plans. It was instantiated with gpt3.5-turbo in a 25-agent sandbox. These are the authors' reported implementation values, not a production game recipe.
• articleWhat could an agent use to check a long draft or a series against its own established facts, and which parts are measured results versus proposals?
  Measured: LLM checkers can find injected contradictions in short stories at roughly human level for the best model and drop toward chance beyond about 1,200-4,000 words (FlawedFictions); a four-stage extract-pair-evidence checker with released prompts found 55% of planted errors in about 8,000-10,000-word stories (ConStory). A 2023 KG+GNN approach on synthetic negation did not beat humans. Proposed only: an explicit fact ledger that is filled from earlier text and queried when new text is written; no inspected paper tests it. This candidate separates the two and defines an experiment.
• articleWhich checks on LLM-written game or story state are actually supported by published tests, and which are only suggested?
  Four checks have a published test in the sources read: (1) rules in the prompt (Wang: game progress 61.5% to 92.1%); (2) deterministic script checks of event conditions and variable updates (RPGBench; they measured errors but did not test whether adding the checker as a retry loop reduced them); (3) a BFS reachability check on a generated game definition (RPGBench game creation; 125 of 500 LLM-generated games valid); (4) an LLM auditor with an explicit fact ledger (NCP-Bench; measures failure, auditor shifts progress numbers). Giving the narrator an explicit state table, and a validator-in-the-loop retry for narrative, are proposals; the only validator-corrected pipeline is Guan's PDDL (planning, not narration).
• articleHow often do LLM narrators contradict established story facts or plot commitments over many turns, and what checks did the benchmark use?
  NCP-Bench (arXiv 2608.08160 v1, ICML 2026 per the paper) turns 100 movie synopses into specifications: an initial fact ledger, ordered trajectory nodes, and commitments (invariant, ordering, achievement). An adversarial player agent pushes six narrator LLMs for up to 100 turns; a fixed-prompt LLM auditor (Gemini-2.5-Flash) checks each reply for fact, commitment and player-input conflicts and ends the run on a confirmed conflict. Best model GPT-5.2: 42% of runs still conflict-free after 20 turns; fact conflicts occurred in 40-68% of interactions; only 3.5% of all 600 runs lasted 100 turns. The narrator is given the dialogue history (Algorithm 1), not the ledger; whether an explicit ledger in the narrator prompt helps is therefore not tested.
• articleIf an LLM updates the numbers (state variables) of a game or story each turn, how often did it get the update wrong in published tests?
  RPGBench (arXiv 2502.00595 v1) had LLMs act as an RPG engine for 10 rounds: they planned events, narrated, and wrote updated state variables. Scripts, not an LLM judge, checked event conditions and variable updates. The best model, Gemini 2.0 Flash Exp, had 76.5% of rounds free of mechanical errors; GPT-4o 69.3%; the lowest were 11.3-20.4% (Claude 3.5 Sonnet, GPT-4o mini, Llama 3.1/3.3 70B). Variable-update error rates per variable ran from 3.4% to 30.8%. Accuracy fell as rounds increased for GPT-4o (69.3% at 10 rounds to 66.8% at 25). Combined with Wang et al. (-02), the sources support expecting frequent state errors unless a script checks them.
• articleWhen an LLM writes several characters' lines, how can I choose among candidates so characters stay distinct from each other?
  TalkMatrix (arXiv v1, 20 Jul 2026) generates k=16 candidate lines per character-situation cell with Qwen3.6-27B-FP8 (Japanese) and picks one per cell jointly to maximise the worst of four embedding-based objectives, one of which is character diversity (mean pairwise cosine distance between different characters' lines in a situation, using a content embedding plus a style embedding). On 25 curated board-game scenarios an LLM judge (GLM-4.7-AWQ) scored the minimax configuration +0.98 over random (6.44 vs 5.46, 17 wins/3 ties/5 losses); independent per-cell reranking by character consistency gave +0.02. Claims are limited to Japanese role-playing scenarios, a model judge and no human ratings.
• articleGive me a step-by-step check for whether a draft's unreliable narration (prose, voice-over or flashback) is fair to the audience.
  No inspected source supplies a validated fairness test. This candidate assembles an eight-step agent check from the sources (labels, ground truth table, signal audit, rival-explanation audit, image-lie audit, hindsight audit, role-stereotype audit, fact audit) and marks each step with its source support. It is a synthesis proposal; the thresholds are not sourced. LLM classifiers for unreliability are reported to be weak outside within-narration tics, so an agent should do the audit with explicit tables and use a model label only as a hint.
• articleHow can an agent measure whether a mystery's reveal is both surprising and fair (set up so the clues add up in hindsight), and what do the open metrics and their human check actually show?
  Wagner, Keydar and Abend (arXiv v2, May 2026) define surprise as how badly a "gullible" reader (an LLM told to take events at face value) predicts the culprit paragraph by paragraph, and coherence as how well clue-competent readers predict it, and argue theoretically that fair play needs a hindsight gap. On whodunit stories their metrics rank Christie above Conan Doyle for surprise and fair play, and LLM-generated mysteries rarely reach both; a 20-person human check gives small correlations (r about 0.1 to 0.25). It is a usable check design for detective outlines and a research result, not a validated test, and code and data were not yet released.
• articleWhich textual signals make a reader (or a classifier) conclude that a narrator is unreliable, and how do the signals differ by what they depend on?
  The Handbook entries list concrete signals: contradiction with real-world fact, mismatch between an earlier report and the actual outcome, other characters' consistent disagreement, internal contradictions and verbal tics, conflicts between narrators, and paratextual types. They also stress that unreliability is a reader's hypothesis competing with other explanations (author error, genre, fictional-world rules, function), so a signal works only if it makes the narrator-based explanation the best one. An open 2025 LLM study operationalises three Hansen types and reports that detection works best for within-narration tics and poorly for signals needing comparison or outside knowledge.
• articleCan an LLM serve as a judge of story or creative-writing quality, and how reliable and biased is it?
  Two 2023 papers disagree. Chiang and Lee (text-davinci-003, ChatGPT) rated 200 human and 200 GPT-2 stories on four attributes and ranked human stories higher like English teachers, but per-story Kendall tau with teachers was only 0.12-0.43 and T0/text-curie-001 did not separate them. Chakrabarty et al. had GPT-3.5, GPT-4 and Claude v1.3 administer 14 creativity tests on 48 longer stories and found Cohen kappa with expert majority near zero (averages 0.016, 0.035, -0.006). Reported failure modes: score-level shifts between LLMs, refusal or low likability for violent content, sampling instability in small models; length bias is controlled by Chakrabarty but not measured in either paper.
• articleWhat is Dramatron, how does it generate a script, and what did 15 theatre and film professionals actually use it for?
  Dramatron (DeepMind, Sept 2022) chains prompts hierarchically from a writer-supplied log line to title, characters, plot (scene beats), location descriptions and per-scene dialogue, using Chinchilla 70B with few-shot prompts; the writer can regenerate, continue or edit at every level. In a 2-hour paid study with 15 playwrights/screenwriters, writers used it mainly for inspiration, character/world alternatives and rough material, not finished scripts; five co-written scripts were staged at the 2022 Edmonton Fringe.
• articleWhat does the WGA New Media Sideletter say about writing credits, and does it mention AI or generative material?
  The New Media Sideletter (revised as of 25 September 2023) has screen-credit paragraphs for derivative new media and high-budget SVOD but no AI wording. Its silence says nothing about other agreements; the 2026 MOA's GAI changes are in Article 72.G.
• articleDo the WGA New Media credits materials (procedures guide and notice form) mention AI or generative material?
  The New Media Credits Procedures Guide (revised October 2020) has no AI wording and treats source material as previously exploited or non-Guild material, so the AI rule comes from the AI page and MOA.
• articleDo the WGA Screen Credits Manual and the Theatrical Credits Procedures Guide say anything about AI or generative material, and where does WGA state the AI credit rule?
  The Screen Credits Manual (2021) has no AI wording; the AI credit rule comes from the WGA AI page and MOA Article 72.C.3.
• articleIf I need to know whether WGA television credit rules address AI-generated material, does the TV Credits Procedures Guide or Manual say anything, and where is the AI rule stated instead?
  The TV credits manual and procedures guide give process and definitions, with no AI wording. The WGA AI page (updated 18 December 2025) states the credit rule: unpublished GAI material is neither assigned nor source material and can't defeat separated rights. Read the manual for process, and the AI page and MOA for AI.
• articleWhat do the Academy Nicholl Fellowships rules require of a feature screenplay entry (format, length, originality, AI use, number of entries), and what differs between the 2023 and 2024 rules?
  The 2024 Nicholl rules require one original English-language feature screenplay as a PDF in standard format (12 pt Courier), 70-160 pages (80-125 suggested), not based on other material. Scripts with AI-generated dialogue, characters or scene description are ineligible, and the entrant earnings cap is US$25,000. Compared with 2023, 2024 adds the AI ineligibility sentence, bans multi-part scripts (2023 allowed stand-alone parts) and no longer carries the 2023 hired-writer sentence in the extracted text. Both years say the Academy acquires no rights in the entry.
• articleWhat evidence exists on the limits or evaluated performance of storylet systems, beyond designers' opinions?
  Evaluated evidence on storylet design is thin: Short (2019) states a structural cost (time caves are the most annoying structure to build with storylets because every piece of content depends on the exact route taken); the LLM storylet framework Drama Llama reports only a six-participant preliminary authoring study (CC BY-NC-SA 4.0 paper).
• articleHas Ofcom published guidance or a decided finding on AI-generated footage in factual programmes, and does it require labelling?
  Ofcom's 3 April 2023 note to broadcasters treats synthetic media under the existing Code (Sections Two, Five, Seven and Eight), warns of harm if audiences cannot tell footage is synthetic, and asks licensees to review compliance; it imposes no labelling requirement, and no decided AI-footage finding was found in the sources read.
• articleWhat does the Writers' Guild of Great Britain publish about generative AI for writers: policy positions, risk guidance for using AI tools, and what it says it has done?
  The Writers' Guild of Great Britain publishes a policy position (survey of 500+ writers; 80% agree developers should seek permission; recommendations on permission, input transparency, payment, output labelling and regulation) and a 2024 guideline warning writers that using generative AI tools carries risks (copyright claims against developers, uncertain status of outputs, breaching contract warranties of originality), marked as a general guide only. Unlike WGA Article 72, these are positions and guidance, not contract clauses.
• articleWhat do the platforms' own creator/advertiser pages say about hooks, pacing and opening seconds for vertical short video, and what do they actually show as evidence?
  TikTok's advertiser best-practice page recommends a hook/body/close structure and states that 90% of ad recall impact is captured in the first six seconds, citing an internal TikTok Creative Guide (2020) without methods on the page.

More