Truace
Subscribe
The Index recomputed live from the record

What the evidence says.What the public feels.

The record holds 935 sourced gains and 771 sourced problems, averaging 68 and 66 on the index score. Readers have logged 22 public signals on the Pulse, which is kept apart and never counted as evidence.

935 gains
771 problems
Every sourced claim in the Index, one square each, shaded by the strength of its evidence. High Moderate Emerging

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

1,706 results
Show filters and sorting

Download every matching row, not just this page:Export CSVExport JSON

AI gains · 935

317
PolicyNewly addedModerate evidence · 1 source

Monthly hydro-climatic machine learning models reproduced observed storage variability at two Taiwan reservoirs with high Nash-Sutcliffe efficiency when static capacity curves were assumed.

By September 2026, a peer-reviewed study tested monthly hydro-climatic machine learning forecasts of effective storage at Mingde and Shihmen reservoirs in Taiwan, comparing linear regularized regression and a nonlinear tree-based model against observed storage and multi-year bathymetric capacity changes.

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

Updated Sep 30, 2026 · TRV-2026-1217

68Index score
318
ScienceNewly addedModerate evidence · 1 source

AI tools now span molecular design, synthesis planning, catalyst optimization, and product verification to compress candidate spaces and accelerate decision-making in organic synthesis workflows.

By September 2026, this peer-reviewed synthesis review found AI in organic chemistry shifting from isolated demonstrations to a deployable toolkit covering design, retrosynthesis planning, condition optimization, and spectroscopic verification, organized as assistant, analyst, and emerging researcher levels.

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

Updated Sep 30, 2026 · TRV-2026-1216

68Index score
319
HealthNewly addedModerate evidence · 1 source

Weighted balanced accuracy (WBAn) provides a regulatory-aligned, risk-based evaluation for multiclass medical AI that weights errors by severity and links development to real-world performance, demonstrated in X-ray lung disease detection.

Researchers propose weighted balanced accuracy (WBAn) to evaluate multiclass machine learning models used in medical devices, aiming to align model assessment with regulatory risk management. They demonstrate the metric using X-ray based detection of lung diseases as a reference scenario.

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

Updated Sep 30, 2026 · TRV-2026-1215

68Index score
320
EducationStableModerate evidence · 1 source

Among surveyed university design students in Pakistan, AI design tools were widely adopted for idea generation and were perceived by a large majority as enhancing originality and productivity.

A peer-reviewed study published October 16, 2025 surveyed 94 university-level art and design students across Pakistan about AI adoption. It found growing use of Midjourney, Adobe Firefly, and ChatGPT primarily for idea generation and style exploration, with 89.36% perceiving AI as enhancing originality and 79.79% viewing AI use in design education positively.

Impact 30%
49
Evidence 25%
95
Scale 20%
35
Confidence 15%
87
Recency 10%
97

Updated Sep 28, 2026 · TRV-2026-1208

68Index score

AI problems · 771

317
HealthStableModerate evidence · 1 source

Artificial intelligence approaches for CT-based prediction of hematoma expansion and adverse outcomes after spontaneous ICH are limited by small cohorts, overfitting, dataset heterogeneity, insufficient external validation, poor interpretability and lack of workflow integration.

This peer-reviewed review in GeroScience appraises CT-based prediction of hematoma expansion after spontaneous intracerebral hemorrhage, a major determinant of early deterioration. It compares contrast-enhanced signs like spot, leakage and iodine signs with non-contrast signs including blend, black hole, island, satellite, hypodensity and swirl signs, plus shape and heterogeneity, and evaluates composite scores and AI approaches including radiomics, machine learning and deep learning.

Impact 30%
49
Evidence 25%
95
Scale 20%
35
Confidence 15%
87
Recency 10%
89

Updated Aug 17, 2026 · TRV-2026-0810

67Index score
318
HealthStableModerate evidence · 1 source

AI adoption in nursing risks replacing relationship-based person-centred fundamental care with algorithm-based care and causing professional infantilisation and loss of critical thinking.

On 2026-08-15 a discursive paper in Journal of Advanced Nursing explored post-humanism and AI in nursing and healthcare through critical reflection on contemporary and established literature. It concluded AI presents both potential benefit and severe threat to fundamental nursing care defined by patient/nurse relationships and patient-centredness carried out with critical reasoning.

Impact 30%
49
Evidence 25%
95
Scale 20%
35
Confidence 15%
87
Recency 10%
89

Updated Aug 17, 2026 · TRV-2026-0809

67Index score
319
EducationStableModerate evidence · 1 source

Despite belief AI will revolutionise nursing education, actual implementation remains conservative at augmentation level with none achieving transformative redefinition.

A scoping review published August 15, 2026 examined how nursing academics perceive and use AI in nursing education, synthesizing 15 studies from eight countries with 2004 academics. It found most believe AI will revolutionise education but actual use is selective and conservative, concentrated at the augmentation level for productivity and research writing rather than assessment or transformative pedagogy.

Impact 30%
49
Evidence 25%
95
Scale 20%
35
Confidence 15%
87
Recency 10%
89

Updated Aug 17, 2026 · TRV-2026-0808

67Index score
320
HealthStableModerate evidence · 1 source

ML applications in clinical microbiology face typical challenges including class imbalance, limited generalization, robustness issues, and need for model interpretation and explainability to achieve robust performance under real-world variability.

Published August 15, 2026, this narrative review in Clinical Microbiology and Infection introduces the machine learning lifecycle from a clinical microbiology perspective, covering data preparation, model development, evaluation, and deployment, drawing on applied research and AI development guidelines for healthcare.

Impact 30%
49
Evidence 25%
95
Scale 20%
35
Confidence 15%
87
Recency 10%
89

Updated Aug 17, 2026 · TRV-2026-0800

67Index score

Recomputed live from the record · Oct 12, 2026, 4:52 PM