Truace
Subscribe
The Index recomputed live from the record

What the evidence says.What the public feels.

The record holds 939 sourced gains and 772 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.

939 gains
772 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.

1,711 results
Show filters and sorting

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

AI gains · 939

513
HealthStableModerate evidence · 1 source

CURL-AID automated analysis of parasternal long-axis echocardiograms quantified posterior annular hypermobility and classified posterior systolic curling with AUC 0.865 and accuracy 0.82.

On 2026-08-12, a peer-reviewed study described CURL-AID, a fully automated framework that segments the posterior mitral annulus and left ventricular wall on parasternal long-axis echocardiograms, tracks tissue motion, and classifies posterior systolic curling using nine kinematic features in 100 retrospectively analyzed patients.

Impact 30%
49
Evidence 25%
95
Scale 20%
35
Confidence 15%
87
Recency 10%
88

Updated Aug 14, 2026 · TRV-2026-0759

67Index score
514
HealthStableModerate evidence · 1 source

AI and machine learning models are being applied to predict chemical toxicity and flag hazardous compounds to inform regulatory decisions, with XAI methods like SHAP and LIME improving interpretability.

On August 12, 2026, a review in the International Journal of Toxicology summarized AI and machine learning use in toxicological risk assessment to predict chemical toxicity and support regulatory decisions, noting that explainable AI methods like SHAP and LIME are being explored to make model decisions interpretable.

Impact 30%
49
Evidence 25%
95
Scale 20%
35
Confidence 15%
87
Recency 10%
88

Updated Aug 14, 2026 · TRV-2026-0757

67Index score
515
ScienceStableModerate evidence · 1 source

Researchers built a DPMD-based machine learning force field with active learning for FLiBeU fuel salt and used it to systematically compute microstructure, thermophysical and transport properties across 773-1173 K and 3-50 mol% UF4.

Researchers developed a high-precision machine learning force field for the molten salt reactor fuel salt LiF-BeF2-UF4 using deep potential molecular dynamics combined with active learning, validated against density functional theory and experiments. They then used the model to systematically calculate microstructural metrics and thermophysical and transport properties across a wide temperature range of 773-1173 K and UF4 concentrations of 3-50 mol%.

Impact 30%
49
Evidence 25%
95
Scale 20%
35
Confidence 15%
87
Recency 10%
88

Updated Aug 14, 2026 · TRV-2026-0756

67Index score
516
HealthStableModerate evidence · 1 source

Additional sectioning at 150 μm intervals with deep learning support detected isolated STIC or HGSC in BRCA1/2 carriers whose initial RRSO pathology was negative but who later developed peritoneal HGSC.

By August 2026, a Histopathology study re-examined fallopian tube tissue from 19 BRCA1/2 carriers who had undergone risk-reducing salpingo-oophorectomy around age 40. Using deeper sections cut at 150 μm intervals and a deep learning model to support STIC detection, the team found occult STIC or HGSC in all patients who later developed peritoneal HGSC despite having no STIC or HGSC at initial diagnosis.

Impact 30%
49
Evidence 25%
95
Scale 20%
35
Confidence 15%
87
Recency 10%
88

Updated Aug 14, 2026 · TRV-2026-0755

67Index score

AI problems · 772

513
HealthStableModerate evidence · 1 source

AI struggles to generate novel hypotheses and apply creative, contextual reasoning, limiting its utility in rare disease diagnosis and complex uncertain scenarios.

A 2026 narrative review and conceptual analysis in Diagnosis synthesized literature on noise in medical decision-making, AI applications in healthcare, and clinical reasoning, reviewing case studies in radiology and pathology and empirical data on AI performance.

Impact 30%
49
Evidence 25%
95
Scale 20%
35
Confidence 15%
87
Recency 10%
82

Updated Jul 14, 2026 · TRV-2026-0218

67Index score
514
ScienceStableModerate evidence · 1 source

Deep neural networks are unable to learn multiple tasks sequentially, suffering catastrophic forgetting when trained on new tasks.

On March 14, 2017, authors in PNAS described a method to overcome catastrophic forgetting in deep neural networks. They noted that while deep networks were the most successful technique for translation, image classification and generation, they could not learn multiple tasks sequentially. Their solution protects weights important for previous tasks, inspired by synaptic consolidation.

Impact 30%
49
Evidence 25%
95
Scale 20%
35
Confidence 15%
87
Recency 10%
82

Updated Jul 13, 2026 · TRV-2026-0210

67Index score
515
PolicyStableModerate evidence · 1 source

The same survey found overreliance on labels, making participants more susceptible to false claims with human-made images and more hesitant to believe true claims illustrated with labeled AI-generated images.

Researchers studied whether legally mandated disclosure labels help users avoid deception from AI-generated images. After five focus groups, they surveyed 1,354 participants on how labels changed their judgments of true and false claims illustrated with different image types.

Impact 30%
49
Evidence 25%
95
Scale 20%
35
Confidence 15%
87
Recency 10%
82

Updated Jul 13, 2026 · TRV-2026-0198

67Index score
516
CrimeStableModerate evidence · 1 source

Generative AI increases the scale and speed of Trust and Safety attacks and lowers barriers to creating sophisticated propaganda and deepfakes.

On April 13, 2026, a peer-reviewed CHI paper reported a qualitative study of 43 Trust & Safety experts across child safety, election integrity, hate and harassment, scams, and violent extremism. It found generative AI both expands attacker capabilities and offers new defensive tools for detection and mitigation.

Impact 30%
49
Evidence 25%
95
Scale 20%
35
Confidence 15%
87
Recency 10%
82

Updated Jul 13, 2026 · TRV-2026-0197

67Index score

Recomputed live from the record · Oct 13, 2026, 3:57 PM