TruaceTracing the truth around AITuesday, September 15, 2026
The Index

What the evidence says.What the public feels.

Ranks distinct AI gain and problem claims from the published record. Scores reward impact, independent source strength, scale, confidence, and recency.

1,419 results
Show filters and sorting

AI gains · 788

74
GainHealth· Newly added· Evidence: Moderate (1 source)

Artificial intelligence-assisted histopathological diagnosis of endocervical gastric-type adenocarcinoma: a multicenter model development and validation study: In internal validation (Stage I), it achieved an accuracy of 0.980 (95% CI 0.977-0.983) and an ROC-AUC of 0.995 (95% CI 0.994-0.997).

Endocervical gastric-type adenocarcinoma (GAS) is one of the most aggressive subtypes of cervical cancer and is frequently underdiagnosed due to morphological ambiguity, leading to delayed diagnosis. Despite the availability of molecular and genomic assays, their high cost, complexity, and limited reproducibility restrict clinical use.

Impact 30%69
Evidence 25%95
Scale 20%35
Confidence 15%87
Recency 10%97

Updated Sep 2, 2026 · TRV-2026-0962

74
GainOther· Stable· Evidence: Moderate (1 source)

Wearable PLF sensors achieved up to 85% accuracy for lameness detection and over 97% accuracy for stress-related physiological changes in extensive sheep systems.

This peer-reviewed review from September 2026 synthesized more than 35 studies on precision livestock farming for extensive sheep production, where large grazing areas and limited supervision make early health detection difficult. It assessed wearable sensors, GPS collars, computer vision, AI and environmental monitoring including drones and ground-based sensors.

Impact 30%69
Evidence 25%95
Scale 20%35
Confidence 15%87
Recency 10%97

Updated Sep 1, 2026 · TRV-2026-0951

74
GainHealth· Stable· Evidence: Moderate (1 source)

Deep learning models for Hirschsprung disease histopathology achieved over 90% ganglion cell detection and cut diagnostic time by 50-95%, increasing accuracy and accelerating clinical decision-making.

This PRISMA 2020 systematic review of thirteen studies from 2016-2025 examined machine and deep learning for Hirschsprung disease diagnosis from histopathological images, evaluating architectures, workflows, and performance with QUADAS-AI and PROBAST.

Impact 30%69
Evidence 25%95
Scale 20%35
Confidence 15%87
Recency 10%97

Updated Aug 31, 2026 · TRV-2026-0942

74
GainHealth· Stable· Evidence: Moderate (1 source)

Automated pulmonary AI matched persisting lung nodules across 3-month LDCT scans with 83.5% success, reaching 91.8% for single-nodule cases and leaving only 1.5% of persisting findings needing manual correction, indicating potential to reduce manual tracking workload.

Researchers tested a pulmonary AI system for fully automated longitudinal nodule matching in 361 UK Lung Cancer Screening trial participants who had 3-month follow-up low-dose CT. Using a >=100 mm3 solid-component threshold, the AI found 378 baseline nodules in 181 participants; 39 resolved, and it matched 283 of 339 persisting nodules for an 83.5% success rate, with 91.8% success in single-nodule cases and 72.8% when more than five nodules were present.

Impact 30%69
Evidence 25%95
Scale 20%35
Confidence 15%87
Recency 10%97

Updated Aug 31, 2026 · TRV-2026-0933

AI problems · 631

72
ProblemBusiness· Stable· Evidence: Moderate (1 source)

Those TFP gains are likely exaggerated and even more modest, predicted to be less than 0.53% over 10 years, because future AI effects will involve hard-to-learn tasks with many context-dependent factors and no objective outcome measures.

This peer-reviewed paper models AI's macroeconomic impact as task-level automation and complementarity, using Hulten's theorem to translate the fraction of tasks impacted and average cost savings into GDP and TFP effects. Using existing exposure estimates, it calculates no more than a 0.66% TFP increase over 10 years, then revises down to less than 0.53% after accounting for the shift from easy-to-learn to hard-to-learn tasks.

Impact 30%63
Evidence 25%95
Scale 20%35
Confidence 15%87
Recency 10%89

Updated Jul 20, 2026 · TRV-2026-0378

72
ProblemLabor· Stable· Evidence: Moderate (1 source)

Human contributions were invisibilised in AI-foregrounded products, with potential displacement in ideation and persistent deskilling and precarious flexible employment for small creative firms.

Published October 28 2024, this peer-reviewed study examined 6 commercial products using AI in creative industries to assess labor market effects. It found AI products were more labor intensive than traditional media because they required both traditional production skills and new computational expertise, while also enabling broader exploration in the ideation phase.

Impact 30%49
Evidence 25%95
Scale 20%60
Confidence 15%87
Recency 10%89

Updated Jul 20, 2026 · TRV-2026-0350

72
ProblemHealth· Stable· Evidence: Moderate (1 source)

Among the same respondents, 70.8% cited technical reliability and 68.2% cited data privacy as top concerns about AI application for cervical screening, while 41.2% reported anxiety during result waiting and 58.1% struggled with medical terminology.

On July 10 2026, a peer-reviewed cross-sectional study reported results from 308 online questionnaire responses about cervical HPV screening experiences. Most respondents were urban women aged 25-35, 76.30% reported a history of HPV infection, and 91.56% had undergone TCT. The study measured current distress points and attitudes toward AI-assisted diagnosis.

Impact 30%63
Evidence 25%95
Scale 20%35
Confidence 15%87
Recency 10%89

Updated Jul 20, 2026 · TRV-2026-0329

72
ProblemHealth· Stable· Evidence: Moderate (1 source)

The same e-CTA tool showed progressively lower sensitivity for more distal occlusions, dropping to 73% for distal M1 with only moderate agreement with experts, and excluded non-target occlusions from primary analysis, requiring adjunctive rather than standalone use.

Between May 2023 and May 2025, researchers retrospectively evaluated 531 multiphase CTA examinations from consecutive patients with suspected acute ischemic stroke at a single center, comparing Brainomix e-CTA automated LVO detection to expert neuroradiologist interpretation.

Impact 30%63
Evidence 25%95
Scale 20%35
Confidence 15%87
Recency 10%89

Updated Jul 20, 2026 · TRV-2026-0307

Recomputed live from the record · Sep 15, 2026, 4:05 PM