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

68
GainHealth· Newly added· Evidence: Moderate (1 source)

Cases without consensus were excluded from the AI accuracy analysis. AI accuracy was calculated as the proportion of correct classifications among consensus cases, and the two models were compared using McNemar's test with continuity correction.

Purpose The purpose of this study was to evaluate the performance of two general-use artificial intelligence models, ChatGPT and Grok, in classifying impacted mandibular third molars using the Pell and Gregory system on panoramic radiographs, compared with a resident consensus reference standard. Materials and methods One hundred panoramic radiographic images of impacted mandibular third molars were independently classified by two blinded resident reviewers using the Pell and Gregory classification system.

Impact 30%49
Evidence 25%95
Scale 20%35
Confidence 15%87
Recency 10%98

Updated Sep 5, 2026 · TRV-2026-0985

68
GainHealth· Newly added· Evidence: Moderate (1 source)

A 17-gene diagnostic signature showed high accuracy for MDS detection (AUC = 0.963). Functional enrichment, machine learning (logistic regression, support vector machine, LASSO), regulatory network (transcription factors, miRNA, RNA-binding proteins, and drug targets), immune infiltration characterization (ssGSEA), and RT-qPCR assessment were performed.

Objective This study aimed to develop a mitochondrial and hematopoiesis-related differentially expressed genes (MH-related DEGs) signature for Myelodysplastic syndromes (MDS) diagnosis and to characterize its regulatory network and immune microenvironment. Methods MH-related DEGs were defined as the intersection of differentially expressed genes from three integrated microarray datasets (GSE145733, GSE19429, GSE81173) with a curated set of mitochondrial- and hematopoiesis-related genes from public databases.

Impact 30%49
Evidence 25%95
Scale 20%35
Confidence 15%87
Recency 10%98

Updated Sep 4, 2026 · TRV-2026-0980

68
GainLifestyle· Newly added· Evidence: Moderate (1 source)

Explainable and domain-adaptive prediction models for refrigerant charging in air conditioning systems within industrial processes: Validation on six AC sensor datasets demonstrates that AC-RPX achieves state-of-the-art accuracy and F1 scores, significantly outperforming conventional deep learning and domain adaptation baselines.

Accurate prediction of refrigerant deficiency in consumer air conditioning (AC) systems is critical for optimizing energy efficiency and operational stability. However, existing data-driven models often suffer from significant performance degradation due to domain shift across different AC types and a lack of explanatory transparency.

Impact 30%49
Evidence 25%95
Scale 20%35
Confidence 15%87
Recency 10%98

Updated Sep 3, 2026 · TRV-2026-0968

68
GainHealth· Newly added· Evidence: Moderate (1 source)

Methodological Reporting Quality of Artificial Intelligence-Based Orthopedic Studies: A Literature Analysis: Work in this area would benefit from closer attention to generalizability, clinical relevance and the existing reporting guidelines.

Objective We set out to examine how completely artificial intelligence (AI)-based orthopedic studies report their methods. Methods A PubMed search covering 2023-2025 was run with a predefined strategy.

Impact 30%49
Evidence 25%95
Scale 20%35
Confidence 15%87
Recency 10%98

Updated Sep 3, 2026 · TRV-2026-0967

AI problems · 631

68
ProblemHealth· Stable· Evidence: Moderate (1 source)

AI-enabled sleep estimation tools raise concerns about racial bias in pulse oximetry, regulatory gaps, variable accuracy, and privacy, requiring further validation to ensure equitable and reliable clinical use.

By August 2026, peer-reviewed guidance for neurologists described a shift in obstructive sleep apnea diagnosis from in-laboratory polysomnography alone to home testing augmented by wearables, nearables, and FDA-cleared software-as-a-medical-device platforms that leverage artificial intelligence and multisignal integration to estimate sleep parameters. The article framed these tools as improving accessibility for patients unable or unwilling to undergo lab studies.

Impact 30%49
Evidence 25%95
Scale 20%35
Confidence 15%87
Recency 10%92

Updated Aug 4, 2026 · TRV-2026-0644

68
ProblemClimate· Stable· Evidence: Moderate (1 source)

The same presence-prediction model failed to identify 2 of 5 APP-contaminated field sites and could not predict particle concentration with sufficient accuracy.

On 2026-08-03, a peer-reviewed study reported a random forest model trained on 16S rRNA gene sequencing from a field mesocosm to predict antifouling paint particle contamination in sediment. In lab incubation tests it identified 100% of presence samples and 83.3% of absence samples, and when applied to 14 Baltic Sea and Warnow estuary sites it correctly labeled all uncontaminated sites and three of five contaminated sites.

Impact 30%49
Evidence 25%95
Scale 20%35
Confidence 15%87
Recency 10%92

Updated Aug 4, 2026 · TRV-2026-0643

68
ProblemHealth· Stable· Evidence: Moderate (1 source)

In prostate cancer and MRI subgroups, AI models showed lower sensitivity than clinicians, indicating inconsistent advantage across tasks.

A meta-analysis of 110 studies up to June 2026 evaluated AI algorithms for diagnosing urological cancers on CT, MRI, and ultrasound, pooling sensitivity, specificity, and AUC and comparing to clinician performance where reported.

Impact 30%49
Evidence 25%95
Scale 20%35
Confidence 15%87
Recency 10%92

Updated Aug 4, 2026 · TRV-2026-0640

68
ProblemHealth· Stable· Evidence: Moderate (1 source)

Deep learning models for knee osteoarthritis progression showed limited generalizability, with performance degradation on external validation and heavy reliance on a single training dataset without rigorous multi-site validation.

This PRISMA systematic review evaluated 33 peer-reviewed studies (2019-2026) comprising 50 deep learning models that predict knee osteoarthritis progression from medical imaging. It extracted AUC as primary outcome, categorized nine different progression definitions, and assessed bias with PROBAST-AI, finding median internal AUCs of 0.87 for surgery, 0.78 for structural, and 0.79 for symptomatic endpoints.

Impact 30%49
Evidence 25%95
Scale 20%35
Confidence 15%87
Recency 10%92

Updated Aug 4, 2026 · TRV-2026-0638

Recomputed live from the record · Sep 15, 2026, 11:37 PM