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

493
CrimeStableModerate evidence · 1 source

Machine learning, deep learning and reinforcement learning techniques improve cybersecurity systems' ability to detect and mitigate cyberattacks including malware and intrusions.

Published January 2024, this IEEE Access survey reviews how machine learning, deep learning and reinforcement learning are applied to cybersecurity tasks such as malware detection, intrusion detection and vulnerability assessment, including evaluation of ChatGPT-like tools on both defensive and offensive sides.

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

Updated Aug 16, 2026 · TRV-2026-0792

67Index score
494
HealthStableModerate evidence · 1 source

An AI model trained on surgical video frames achieved automated real-time detection and segmentation of bladder neck, adenoma and peripheral zone during robot-assisted prostate enucleation with high Dice scores and >60 fps inference.

In a retrospective single-centre pilot, investigators developed a YOLOv11-based model to automatically detect and segment three key landmarks during robot-assisted single-port transvesical enucleation of the prostate using 611 annotated frames from 37 procedures performed by one expert surgeon.

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

Updated Aug 16, 2026 · TRV-2026-0791

67Index score
495
HealthStableModerate evidence · 1 source

A zero-shot LLM-assisted workflow improved automated identification of cardiovascular events from EMRs, achieving the highest AUCs for stroke, MI and composite MACE in two cohorts compared to ICD codes, primary diagnosis and problem list methods.

A multisite retrospective validation study in a US tertiary health system compared four automated EMR retrieval methods to manual chart adjudication for ischaemic stroke/TIA, MI, HF exacerbation/hospitalisation and composite MACE in 2258 patients treated with immune checkpoint inhibitors and 1426 patients who underwent TAVR. The zero-shot LLM workflow achieved the highest AUCs for most outcomes, while ICD-based retrieval remained competitive.

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

Updated Aug 16, 2026 · TRV-2026-0790

67Index score
496
HealthStableModerate evidence · 1 source

An AI reinforcement learning model trained on Swedish population data produced efficient, low-cost diagnostic sequences for moderate to severe breathlessness with high diagnostic yield, prioritizing BMI, anxiety/depression, physical activity, and spirometry before lung diffusion, CT, and hemoglobin tests.

On 2026-08-13, researchers reported developing an AI reinforcement learning model using Swedish population data to determine cost-effective diagnostic pathways for chronic breathlessness. The model evaluated 16 clinically relevant conditions with associated tests and costs, generating tailored sequences for subgroups defined by sex and smoking exposure.

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

Updated Aug 16, 2026 · TRV-2026-0789

67Index score

AI problems · 772

493
HealthStableModerate evidence · 1 source

AI integration in radiologic practice may paradoxically increase workload and contribute to radiologist burnout when poorly implemented, with automation bias and over-reliance compromising clinical judgment

This narrative review from June 2026 synthesized evidence on AI integration in radiology, finding that by that date AI systems had shown diagnostic performance approaching or exceeding radiologists in chest imaging and breast cancer screening and had improved triage and reduced report turnaround times in practice.

Impact 30%
49
Evidence 25%
95
Scale 20%
35
Confidence 15%
87
Recency 10%
84

Updated Jul 20, 2026 · TRV-2026-0311

67Index score
494
HealthStableModerate evidence · 1 source

Adoption of AI in fragile humanitarian environments creates substantial risk of security breaches from human errors and unregulated data management, and risks reinforcing existing power imbalances for workers and communities.

A peer-reviewed commentary from April 2026 describes how AI tools are being adopted in the humanitarian cooperation sector to improve health diagnostics, service quality, and efficiency of analysis and data management for emergency responses in conflict areas with limited resources.

Impact 30%
49
Evidence 25%
95
Scale 20%
35
Confidence 15%
87
Recency 10%
84

Updated Jul 20, 2026 · TRV-2026-0310

67Index score
495
HealthStableModerate evidence · 1 source

GPT-Reviewer showed near-zero agreement with human QUIPS ratings for study participation and outcome measurement, with kappa 0.001.

Researchers nested a two-part methodological study within two PROSPERO-registered reviews to test customized GPT models on complex rheumatology evidence synthesis. Fifteen SLE metabolomics studies were used to compare human and GPT data extraction, and nineteen rheumatology prognostic studies were reappraised in 2025 with GPT-Reviewer against adjudicated human QUIPS ratings using weighted kappa.

Impact 30%
49
Evidence 25%
95
Scale 20%
35
Confidence 15%
87
Recency 10%
84

Updated Jul 20, 2026 · TRV-2026-0308

67Index score
496
HealthStableModerate evidence · 1 source

Translational progress of AI applied to bioink formulation and bioprinting for bone regeneration remains constrained by small datasets, limited cross-platform validation, and weak links between computational predictions and biological outcomes.

By July 10 2026, a peer-reviewed review in Tissue Engineering Part B: Reviews evaluated experimentally validated AI uses in scaffold-based bone regeneration, from materials design to fabrication control and biological assessment. It reported that physics-informed models tend to be more robust and generalizable than purely data-driven models, while transfer learning is hampered by variability in cellular responses and fabrication conditions.

Impact 30%
49
Evidence 25%
95
Scale 20%
35
Confidence 15%
87
Recency 10%
84

Updated Jul 20, 2026 · TRV-2026-0288

67Index score

Recomputed live from the record · Oct 13, 2026, 2:08 PM