scan.json (19037B)
1 { 2 "paper": { 3 "title": "Hogyan igazodjunk el a mesterséges intelligencia munkaerőpiaci hatásait övező zajban?", 4 "authors": ["Andrea Szalavetz"], 5 "year": 2026, 6 "venue": "Közgazdasági Szemle", 7 "doi": "10.18414/KSZ.2026.1.72" 8 }, 9 "scan_version": 2, 10 "active_modules": [], 11 "methodology_tags": ["qualitative", "theoretical"], 12 "key_findings": "The paper critically reviews conflicting expert assessments of AI's labor market impact, identifying sources of 'noise' in predictions. It argues the relationship between AI's theoretical capabilities and actual task automation is non-linear. Applying the skill-biased technological change (SBTC) framework to the AI era, it concludes that AI will devalue the qualifications of low-to-medium-skilled graduates, intensifying competition for knowledge-intensive jobs and favoring only the most capable workers.", 13 "checklist": { 14 "artifacts": { 15 "code_released": { 16 "applies": false, 17 "answer": false, 18 "justification": "This is a narrative literature review with no code or computational analysis to release." 19 }, 20 "data_released": { 21 "applies": false, 22 "answer": false, 23 "justification": "No original dataset was collected or used. The paper synthesizes existing literature and anecdotal examples." 24 }, 25 "environment_specified": { 26 "applies": false, 27 "answer": false, 28 "justification": "No computational environment is involved in this theoretical/qualitative literature review." 29 }, 30 "reproduction_instructions": { 31 "applies": false, 32 "answer": false, 33 "justification": "No computational experiments to reproduce. This is a discursive essay." 34 } 35 }, 36 "statistical_methodology": { 37 "confidence_intervals_or_error_bars": { 38 "applies": false, 39 "answer": false, 40 "justification": "No original statistical analysis is performed. The paper cites others' findings narratively." 41 }, 42 "significance_tests": { 43 "applies": false, 44 "answer": false, 45 "justification": "No original statistical analysis or comparative claims based on the author's own data." 46 }, 47 "effect_sizes_reported": { 48 "applies": false, 49 "answer": false, 50 "justification": "No original empirical analysis. Effect sizes cited are from referenced studies, not the author's work." 51 }, 52 "sample_size_justified": { 53 "applies": false, 54 "answer": false, 55 "justification": "No original data collection; this is a literature review essay." 56 }, 57 "variance_reported": { 58 "applies": false, 59 "answer": false, 60 "justification": "No original experiments or quantitative analysis." 61 } 62 }, 63 "evaluation_design": { 64 "baselines_included": { 65 "applies": false, 66 "answer": false, 67 "justification": "No system or method is evaluated. This is a critical essay reviewing the state of discourse." 68 }, 69 "baselines_contemporary": { 70 "applies": false, 71 "answer": false, 72 "justification": "No evaluation or system comparison is performed." 73 }, 74 "ablation_study": { 75 "applies": false, 76 "answer": false, 77 "justification": "No system with components to ablate." 78 }, 79 "multiple_metrics": { 80 "applies": false, 81 "answer": false, 82 "justification": "No evaluation metrics are used; this is a discursive review." 83 }, 84 "human_evaluation": { 85 "applies": false, 86 "answer": false, 87 "justification": "No system outputs to evaluate." 88 }, 89 "held_out_test_set": { 90 "applies": false, 91 "answer": false, 92 "justification": "No datasets or test sets involved." 93 }, 94 "per_category_breakdown": { 95 "applies": false, 96 "answer": false, 97 "justification": "No quantitative results to break down." 98 }, 99 "failure_cases_discussed": { 100 "applies": false, 101 "answer": false, 102 "justification": "No system or method being evaluated." 103 }, 104 "negative_results_reported": { 105 "applies": false, 106 "answer": false, 107 "justification": "No experiments conducted." 108 } 109 }, 110 "claims_and_evidence": { 111 "abstract_claims_supported": { 112 "applies": true, 113 "answer": true, 114 "justification": "The abstract claims are qualitative — reviewing conflicting conclusions, applying SBTC to the AI era, arguing AI devalues graduate qualifications. The body substantiates each of these through literature discussion." 115 }, 116 "causal_claims_justified": { 117 "applies": true, 118 "answer": false, 119 "justification": "The paper makes causal claims ('AI devalues qualifications', 'demand will decline for low-to-medium-skilled graduates') based on theoretical reasoning and anecdotal evidence, not causal identification strategies. The author acknowledges this is speculative ('I assign greater probability to the pessimistic scenario')." 120 }, 121 "generalization_bounded": { 122 "applies": true, 123 "answer": true, 124 "justification": "The author explicitly notes the focus is narrow — primarily US data — and states 'the analysed literature and predictions predominantly draw conclusions from US data' as a limitation. Future work will examine countries outside the most developed ones." 125 }, 126 "alternative_explanations_discussed": { 127 "applies": true, 128 "answer": true, 129 "justification": "The paper is structured around presenting competing explanations. For example, it discusses whether graduate employment decline is due to AI or cyclical/structural factors, citing Eckhardt & Goldschlag, Lettink, Martin, and Smith alongside the AI-causal view." 130 }, 131 "proxy_outcome_distinction": { 132 "applies": false, 133 "answer": false, 134 "justification": "No original measurements are made. The paper is a theoretical/qualitative essay." 135 } 136 }, 137 "setup_transparency": { 138 "model_versions_specified": { 139 "applies": false, 140 "answer": false, 141 "justification": "No models are used by the author. AI models are discussed as subject matter." 142 }, 143 "prompts_provided": { 144 "applies": false, 145 "answer": false, 146 "justification": "No prompting is used in this paper." 147 }, 148 "hyperparameters_reported": { 149 "applies": false, 150 "answer": false, 151 "justification": "No experiments or model usage by the author." 152 }, 153 "scaffolding_described": { 154 "applies": false, 155 "answer": false, 156 "justification": "No agentic scaffolding used." 157 }, 158 "data_preprocessing_documented": { 159 "applies": false, 160 "answer": false, 161 "justification": "No data preprocessing. The paper explicitly states it does not claim to be a systematic review ('módszeres irodalomáttekintés igénye nélkül')." 162 } 163 }, 164 "limitations_and_scope": { 165 "limitations_section_present": { 166 "applies": true, 167 "answer": true, 168 "justification": "The closing section ('Zárógondolatok') explicitly discusses three limitations: the noise mapping doesn't answer the title question, the analysis is a snapshot only 3 years after ChatGPT, and the geographic focus is limited to US data." 169 }, 170 "threats_to_validity_specific": { 171 "applies": true, 172 "answer": true, 173 "justification": "The author identifies specific threats: the geographic bias toward US data, the short time window (33 months since ChatGPT), and acknowledges that 'my own prediction... can rightly be considered merely another opinion adding to the noise.' These are specific to this study." 174 }, 175 "scope_boundaries_stated": { 176 "applies": true, 177 "answer": true, 178 "justification": "The paper explicitly lists topics it did NOT examine: AI's effect on low-skilled non-routine jobs, human-AI collaboration dynamics, policy interventions, and conditions under which AI augments vs. replaces skills. States these are 'topics for future research.'" 179 } 180 }, 181 "data_integrity": { 182 "raw_data_available": { 183 "applies": false, 184 "answer": false, 185 "justification": "No original data collected. This is a literature review essay." 186 }, 187 "data_collection_described": { 188 "applies": true, 189 "answer": false, 190 "justification": "The paper explicitly disclaims systematic literature review methodology ('módszeres irodalomáttekintés igénye nélkül' — without claiming to be a systematic review). No search strategy, databases, or selection criteria are described." 191 }, 192 "recruitment_methods_described": { 193 "applies": false, 194 "answer": false, 195 "justification": "No human participants. No data collection requiring recruitment." 196 }, 197 "data_pipeline_documented": { 198 "applies": false, 199 "answer": false, 200 "justification": "No data pipeline. Literature selection is not documented." 201 } 202 }, 203 "conflicts_of_interest": { 204 "funding_disclosed": { 205 "applies": true, 206 "answer": false, 207 "justification": "No funding source is mentioned. The author is affiliated with ELTE KRTK (a Hungarian research institute), but no grant or funding acknowledgment is provided." 208 }, 209 "affiliations_disclosed": { 210 "applies": true, 211 "answer": true, 212 "justification": "The author's affiliation is clearly stated: 'Szalavetz Andrea DSc, az ELTE KRTK Világgazdasági Intézetének tudományos tanácsadója' (scientific advisor at ELTE KRTK Institute of World Economics)." 213 }, 214 "funder_independent_of_outcome": { 215 "applies": false, 216 "answer": false, 217 "justification": "No funder disclosed. The author is at a public research institute, and the paper takes no position that would create a financial conflict." 218 }, 219 "financial_interests_declared": { 220 "applies": true, 221 "answer": false, 222 "justification": "No competing interests statement is provided." 223 } 224 }, 225 "contamination": { 226 "training_cutoff_stated": { 227 "applies": false, 228 "answer": false, 229 "justification": "No pre-trained model is evaluated on any benchmark." 230 }, 231 "train_test_overlap_discussed": { 232 "applies": false, 233 "answer": false, 234 "justification": "No model evaluation performed." 235 }, 236 "benchmark_contamination_addressed": { 237 "applies": false, 238 "answer": false, 239 "justification": "No benchmark evaluation performed." 240 } 241 }, 242 "human_studies": { 243 "pre_registered": { 244 "applies": false, 245 "answer": false, 246 "justification": "No human participants in this literature review." 247 }, 248 "irb_or_ethics_approval": { 249 "applies": false, 250 "answer": false, 251 "justification": "No human participants." 252 }, 253 "demographics_reported": { 254 "applies": false, 255 "answer": false, 256 "justification": "No human participants." 257 }, 258 "inclusion_exclusion_criteria": { 259 "applies": false, 260 "answer": false, 261 "justification": "No human participants." 262 }, 263 "randomization_described": { 264 "applies": false, 265 "answer": false, 266 "justification": "No human participants." 267 }, 268 "blinding_described": { 269 "applies": false, 270 "answer": false, 271 "justification": "No human participants." 272 }, 273 "attrition_reported": { 274 "applies": false, 275 "answer": false, 276 "justification": "No human participants." 277 } 278 }, 279 "cost_and_practicality": { 280 "inference_cost_reported": { 281 "applies": false, 282 "answer": false, 283 "justification": "Theoretical/qualitative paper with no method to cost." 284 }, 285 "compute_budget_stated": { 286 "applies": false, 287 "answer": false, 288 "justification": "No computation performed." 289 } 290 } 291 }, 292 "claims": [ 293 { 294 "claim": "The relationship between AI's theoretical capabilities and actual automation of tasks is non-linear; once AI performance crosses a task-specific threshold, automation of that task can accelerate suddenly.", 295 "evidence": "Theoretical argument supported by reference to Szalavetz (2019) and discussion of the gap between benchmark performance and real-world deployment. No original empirical evidence provided.", 296 "supported": "weak" 297 }, 298 { 299 "claim": "AI will shift the skill-biased technological change (SBTC) dynamic to operate WITHIN the graduate workforce, devaluing low-to-medium-skilled graduates rather than only affecting non-graduate workers.", 300 "evidence": "Theoretical extension of SBTC framework (Autor et al. 1998, 2003; Goldin & Katz 1998) to AI era, supported by anecdotal evidence from 'AI Killed My Job' blog series (Merchant 2025) and labor market statistics showing declining demand for recent graduates (Brynjolfsson, Chandar et al. 2025; Hosseini & Lichtinger 2025).", 301 "supported": "moderate" 302 }, 303 { 304 "claim": "AI's skill-equalizing effect (boosting low-performers more than high-performers) is only temporary; longer-term, it benefits the most skilled workers while displacing the merely adequate.", 305 "evidence": "Cites Dell'Acqua et al. (2023) and Noy & Zhang (2023) for the short-term equalizing effect, then argues theoretically that once firms recognize AI can handle routine knowledge tasks, they reduce headcount. Supported by anecdotal evidence from the Merchant (2025) blog series. The temporal argument is the author's own inference.", 306 "supported": "weak" 307 }, 308 { 309 "claim": "AI productivity gains are delayed at the macro level due to the need for complementary organizational innovations, retraining, and resource restructuring.", 310 "evidence": "Cites Brynjolfsson et al. (2017), McElheran et al. (2025), Teece (1986) on complementary assets, and Becker et al. (2025) showing AI reduced developer productivity. Also cites Acemoglu (2025) and the productivity paradox literature (David 1990, Van Ark 2016).", 311 "supported": "moderate" 312 } 313 ], 314 "red_flags": [ 315 { 316 "flag": "Not a systematic review", 317 "detail": "The author explicitly disclaims systematic review methodology ('módszeres irodalomáttekintés igénye nélkül'). Literature selection appears ad hoc with no documented search strategy, inclusion criteria, or quality assessment of sources. This risks cherry-picking studies that support the author's narrative." 318 }, 319 { 320 "flag": "Heavy reliance on anecdotal evidence", 321 "detail": "Key claims about AI displacing graduate workers are supported largely by news articles, blog posts (Merchant's 'AI Killed My Job'), and CEO statements (Amodei, Altman), alongside peer-reviewed studies. The anecdotal evidence is presented with similar weight to empirical research." 322 }, 323 { 324 "flag": "Acknowledged opinion without data", 325 "detail": "The author's main prediction ('I assign greater probability to the pessimistic scenario') is explicitly acknowledged as opinion. The paper quotes 'Without data, you're just another person with an opinion' and concedes this applies to its own conclusions." 326 } 327 ], 328 "cited_papers": [ 329 { 330 "title": "The simple macroeconomics of AI", 331 "authors": ["Daron Acemoglu"], 332 "year": 2025, 333 "doi": "10.1093/epolic/eiae042", 334 "relevance": "Influential analysis of AI's macroeconomic productivity impact, argues effects are modest so far." 335 }, 336 { 337 "title": "Generative AI at work", 338 "authors": ["Erik Brynjolfsson", "Danielle Li", "Lindsey Raymond"], 339 "year": 2025, 340 "doi": "10.1093/qje/qjae044", 341 "relevance": "Key empirical study showing generative AI boosts low-skilled worker productivity most, published in QJE." 342 }, 343 { 344 "title": "Canaries in the coal mine? Six facts about the recent employment effects of Artificial Intelligence", 345 "authors": ["Erik Brynjolfsson", "Bhavin Chandar", "Rubing Chen"], 346 "year": 2025, 347 "relevance": "Documents early employment effects of AI on graduate workers, directly relevant to AI labor impact research." 348 }, 349 { 350 "title": "GDPVal: Evaluating AI Model Performance on Real-World Economically Valuable Tasks", 351 "authors": ["Tarun Patwardhan", "Ramana Dias"], 352 "year": 2025, 353 "relevance": "OpenAI study comparing AI vs human performance on 1,320 real-world knowledge tasks across 44 occupations." 354 }, 355 { 356 "title": "The AI Productivity Index (APEX)", 357 "authors": ["Bertie Vidgen"], 358 "year": 2025, 359 "doi": "10.48550/arXiv.2509.25721", 360 "relevance": "Benchmark evaluating AI 'win rates' against human workers across professional tasks." 361 }, 362 { 363 "title": "Navigating the jagged technological frontier: Field experimental evidence of the effects of AI on knowledge worker productivity and quality", 364 "authors": ["Fabrizio Dell'Acqua"], 365 "year": 2023, 366 "doi": "10.2139/ssrn.4573321", 367 "relevance": "Harvard field experiment on AI effects on knowledge worker productivity, foundational study in this space." 368 }, 369 { 370 "title": "Experimental evidence on the productivity effects of generative artificial intelligence", 371 "authors": ["Shakked Noy", "Whitney Zhang"], 372 "year": 2023, 373 "doi": "10.1126/science.adh2586", 374 "relevance": "Science paper with experimental evidence on generative AI productivity effects." 375 }, 376 { 377 "title": "Measuring the impact of early-2025 AI on experienced open-source developer productivity", 378 "authors": ["Joost Becker", "Nicky Rush", "Everett Barnes", "David Rein"], 379 "year": 2025, 380 "relevance": "Study finding AI decreased experienced developer productivity, counter-evidence to productivity gains narrative." 381 }, 382 { 383 "title": "GPTs are GPTs: Labor market impact potential of LLMs", 384 "authors": ["Tyna Eloundou", "Sam Manning", "Pamela Mishkin", "Daniel Rock"], 385 "year": 2024, 386 "doi": "10.1126/science.adj0998", 387 "relevance": "Seminal paper on LLM exposure across occupations, published in Science." 388 }, 389 { 390 "title": "Which economic tasks are performed with AI? Evidence from millions of Claude conversations", 391 "authors": ["Kunal Handa"], 392 "year": 2025, 393 "relevance": "Anthropic study analyzing 4 million Claude conversations to determine AI usage patterns across occupations." 394 }, 395 { 396 "title": "AI as normal technology", 397 "authors": ["Arvind Narayanan", "Sayash Kapoor"], 398 "year": 2025, 399 "relevance": "Argues AI is 'ordinary technology' with non-catastrophic labor effects, directly relevant to AI capability and impact assessment." 400 } 401 ] 402 }