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

345
ScienceStableModerate evidence · 1 source

Deep-learning model automatically segmented 52 anatomical regions on brain FDG PET alone and achieved MRI-comparable pons-based SUVR quantification across cognitively normal and neurodegenerative patients.

By September 2026, a peer-reviewed study in Brain reported development and validation of a deep-learning model that automatically segments 52 anatomical regions on brain [18F]FDG PET images without requiring a paired MRI, tested on 1628 images spanning cognitively normal subjects and patients with mild cognitive impairment, Alzheimer's disease, frontotemporal lobar degeneration and Lewy body dementia.

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

Updated Sep 16, 2026 · TRV-2026-1110

68Index score
346
BusinessStableModerate evidence · 1 source

Machine learning classification improves risk differentiation for electric-vehicle traffic insurance by predicting claim amount levels from policy and behavioral features.

A peer-reviewed study in Traffic Injury Prevention collected electric-vehicle traffic insurance claims from Changsha, China, applied SMOTE for class balancing, and tested six machine learning models to classify Claim Amount Level, using SHAP to explain feature influence.

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

Updated Sep 16, 2026 · TRV-2026-1106

68Index score
347
ScienceStableModerate evidence · 1 source

Life-science physics students engaged in prompting, refining, and validating generative AI-constructed simulations as a new method for virtual learning about physical phenomena.

As of the September 2025 publication date, researchers investigated a novel use of generative AI in physics instruction where students in a second-semester course for life science majors were asked to prompt, refine, and validate AI-constructed simulations of physical phenomena in a lab focused on electric topics, comparing three instructional approaches.

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

Updated Sep 15, 2026 · TRV-2026-1100

68Index score
348
HealthStableModerate evidence · 1 source

Random forest with SHAP and network analysis among 6573 adolescents ranked key contributors to IGD risk and identified central nodes, which may inform early risk identification.

By September 2026, researchers reported using four machine learning algorithms to classify Internet Gaming Disorder risk among 6573 adolescents, selecting a random forest model after comprehensive performance evaluation. They applied SHAP to rank feature contributions and network analysis to map interrelationships among psychosocial factors.

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

Updated Sep 15, 2026 · TRV-2026-1097

68Index score

AI problems · 771

345
Media & ArtsStableModerate evidence · 1 source

Participating writers worried that AI assistance could weaken creative control, interrupt their established process or reproduce a fixed version of their style rather than support continued experimentation and growth.

A mixed-methods study examined how 19 professional writers and 30 avid readers understood authenticity in writing produced with AI assistance. Writers completed short writing tasks using both personalized and non-personalized GPT-4 suggestions, while readers evaluated passages written independently or with either form of AI support.

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

Updated Aug 8, 2026 · TRV-2026-0692

67Index score
346
Media & ArtsStableModerate evidence · 1 source

Fully automated AI generation of news risks bias, manipulation, lack of accountability, inaccuracy, and erosion of accuracy, trust, and copyright principles.

By June 2026, a peer-reviewed paper analyzed how AI and AIGC are being integrated into newsrooms, from data mining to co-creation of news products, increasing efficiency and output volume while prompting questions about human professionalism and editorial control.

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

Updated Aug 8, 2026 · TRV-2026-0697

67Index score
347
PolicyStableModerate evidence · 1 source

Personal-data-driven algorithmic targeting and other manipulative design features like infinite scrolling directly cause primary harms of manipulation, including hijacking attention, overriding user autonomy, and engineering addiction.

Published July 6 2026, this law review article argues that Section 230 and the First Amendment do not categorically immunize digital platforms for harms caused by their own design choices. It proposes a typology separating direct primary harms from design decisions from secondary harms from user content and tertiary harms, focusing on personal-data-driven algorithmic targeting and dark patterns like infinite scrolling.

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

Updated Aug 8, 2026 · TRV-2026-0695

67Index score
348
HealthStableModerate evidence · 1 source

ChatGPT-5.0 did not achieve the highest overall classification accuracy and had lower specificity than XGBoost, indicating it missed the top performance for correctly identifying non-extraction cases.

A comparative study published August 7, 2026 evaluated ChatGPT-5.0 against five supervised machine learning algorithms for orthodontic extraction decisions. Using 520 cases (42.88% extraction, 57.12% non-extraction) and 23 clinical, cephalometric and photographic variables, with expert consensus as reference, ChatGPT-5.0 achieved 75.77% accuracy and 76.68% sensitivity under 5-fold cross-validation, compared to 78.08% accuracy for XGBoost.

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

Updated Aug 8, 2026 · TRV-2026-0689

67Index score

Recomputed live from the record · Oct 12, 2026, 9:28 PM