TruaceTracing the truth around AIWednesday, August 26, 2026

All traces

Comparison of artificial intelligence-based chatbots and expert periodontists in responding to patient questions: a multi-dimensional analysis
HealthContested · G 69 / P 67

LLM chatbots answering periodontal patient questions for patient education

Source article: Comparison of artificial intelligence-based chatbots and expert periodontists in responding to patient questions: a multi-dimensional analysis

Problem

Some chatbot models showed lower conciseness & focus and clarity, producing longer less-focused answers that may make it harder for patients to maintain focus and perceive information in a structured manner.

BMC Oral Health
Gain

LLM chatbots answered periodontal patient questions with scientific accuracy comparable to expert periodontologists while scoring higher on completeness and empathy.

BMC Oral Health
Explainable artificial intelligence techniques for interpretation of food models: a review
LifestyleContested · G 68 / P 67

use of explainable AI techniques to interpret food quality control models for contaminant detection and freshness assessment

Source article: Explainable artificial intelligence techniques for interpretation of food models: a review

Problem

Opaque AI decision-making in food quality models hinders adoption by inspectors and limits reliability, with explainable AI still underutilized in Food Engineering.

Artificial Intelligence Review
Gain

Explainable AI methods such as SHAP and Grad-CAM improve food quality control by identifying which spectral wavelengths or image regions drive predictions, increasing transparency for inspectors verifying contaminant detection and freshness assessments.

Artificial Intelligence Review
AI-Assisted cardiomegaly screening via implicit morphological inference and human-in-the-loop validation
HealthContested · G 70 / P 73

UBNet-Seg cardiomegaly screening accuracy on external NIH and OpenI datasets

Source article: AI-Assisted cardiomegaly screening via implicit morphological inference and human-in-the-loop validation

Problem

Fully automated cardiomegaly screening accuracy dropped to 76.07% on the external OpenI dataset, showing domain-shift vulnerability, while manual CTR measurement remains a clinical bottleneck and existing deep models suffer from algorithmic bloating.

Journal of X-Ray Science and Technology
Gain

UBNet-Seg, a 2.3M-parameter U-Net variant using lung fields as geometric proxy, achieved 95.85% lung Dice at 0.05s inference and 90.31% automated cardiomegaly accuracy on NIH, rising to 93.63% on NIH and 91.21% on OpenI after expert-guided refinement.

Journal of X-Ray Science and Technology
Advancements in machine learning and deep learning for early detection and management of mental health disorder
HealthContested · G 67 / P 69

use of ML and DL for early diagnosis and treatment of mental health disorders

Source article: Advancements in machine learning and deep learning for early detection and management of mental health disorder

Problem

Use of ML and DL for mental health early detection faces challenges with data integration, methodological inconsistency, and ethical issues.

Journal of Affective Disorders Reports
Gain

ML and DL analysis of imaging, genetics, and behavioral data can improve early identification and diagnostic accuracy for conditions like depression, bipolar disorder, and schizophrenia.

Journal of Affective Disorders Reports
Relationships in the age of AI: A review on the opportunities and risks of synthetic relationships to reduce loneliness
LifestyleNegative state · G 64 / P 69

AI companion synthetic relationships for loneliness

Source article: Relationships in the age of AI: A review on the opportunities and risks of synthetic relationships to reduce loneliness

Problem

Widespread use of synthetic AI companions risks emotional over-reliance, distorted expectations for human interaction, privacy harms, and altered norms of intimacy.

Computers in Human Behavior Reports
Gain

Synthetic relationships with AI companions could reduce loneliness by providing always-available, adaptive, emotionally responsive interaction that fosters companionship and lowers social anxiety.

Computers in Human Behavior Reports
Co-designing AI systems with value-sensitive citizen science
PolicyContested · G 66 / P 69

public participation and governance in AI development via Value-Sensitive Citizen Science

Source article: Co-designing AI systems with value-sensitive citizen science

Problem

Existing AI development remains monocultural and top-down, with gaps in inclusion and persistent power asymmetries and epistemic justice concerns.

AI & SOCIETY
Gain

VSCS framework enables community members to act as co-researchers and translate local values into technical requirements for AI systems.

AI & SOCIETY
Algorithmic Justice and Responsible AI Journalism: A Comparative Communication Policy Perspective in East Asia
PolicyContested · G 70 / P 68

regulatory models for AI journalism and algorithmic news curation in China and South Korea

Source article: Algorithmic Justice and Responsible AI Journalism: A Comparative Communication Policy Perspective in East Asia

Problem

The same Chinese and South Korean regulatory models for AI journalism face contrasting trade-offs between regulatory efficiency and editorial independence.

PhilPapers (PhilPapers Foundation)
Gain

Chinese and South Korean regulatory toolkits for AI journalism on platforms like Toutiao and Naver were found to mitigate risks of digital infodemics and algorithmic bias.

PhilPapers (PhilPapers Foundation)
Predicting Biomolecular Interactions in the Next Decade: Physics-Based Methods Meet AI-Driven Approaches
ScienceContested · G 71 / P 70

predicting biomolecular recognition ensembles and thermodynamic observables using machine learning

Source article: Predicting Biomolecular Interactions in the Next Decade: Physics-Based Methods Meet AI-Driven Approaches

Problem

Machine learning models for biomolecular recognition do not inherently enforce thermodynamic consistency and may produce configurations that are not physically realizable.

The Journal of Physical Chemistry Letters
Gain

Data-driven machine learning models can rapidly generate biomolecular structures and propose conformational ensembles for recognition events with high predictive performance.

The Journal of Physical Chemistry Letters
Machine learning force fields for inorganic crystalline materials: principles, advances, and emerging applications
ScienceContested · G 74 / P 71

machine learning force fields for inorganic crystalline materials

Source article: Machine learning force fields for inorganic crystalline materials: principles, advances, and emerging applications

Problem

MLFFs still face challenges in computational efficiency and scale, accuracy and generalization, data requirements, interpretability, and physical constraints when applied to inorganic crystalline materials.

Physical Chemistry Chemical Physics
Gain

MLFFs provide high accuracy with high efficiency for atomic-level studies of inorganic crystalline materials, overcoming traditional limits in structure prediction, properties, defects, and phase transitions.

Physical Chemistry Chemical Physics
Medical student reliance on artificial intelligence in nephrology education
HealthContested · G 72 / P 74

accuracy after reviewing ChatGPT-generated answers during pediatric nephrology case-based learning in first-year medical students

Source article: Medical student reliance on artificial intelligence in nephrology education

Problem

First-year medical students changed answers to match ChatGPT in 22.3% of cases, with greater reliance on foundational than clinical questions, indicating context-dependent overreliance risk.

Journal of Nephrology
Gain

First-year medical students showed a modest net improvement in accuracy after reviewing ChatGPT-generated answers, because incorrect-to-correct changes exceeded correct-to-incorrect changes.

Journal of Nephrology
‘Not up for grabs’: Albanese establishes AI office and vows to protect Australian creatives from copyright ‘theft’
BusinessContested · G 63 / P 67

use of Australian creative works to build and train AI models

Source article: ‘Not up for grabs’: Albanese establishes AI office and vows to protect Australian creatives from copyright ‘theft’

Problem

AI companies have been able to use Australian books, music, art and news to build and train models without artist control or compensation, while communities face impacts from large energy-intensive datacentres competing for land, power and water.

Science and Technology Law Review
Gain

Australian creatives would retain ownership and control of their work and receive payment through licensing deals with AI companies, rather than having their books, music, art and news used for free to train models.

The Guardian
Advancing Decision-Making through AI-Human Collaboration: A Systematic Review and Conceptual Framework
ScienceContested · G 69 / P 72

AI-human collaborative decision-making and how its dynamics and typologies shape decision outcomes

Source article: Advancing Decision-Making through AI-Human Collaboration: A Systematic Review and Conceptual Framework

Problem

Existing literature on AI-human decision-making was fragmented and lacked integrative frameworks to explain how AI-human dynamics and decision typologies shape outcomes despite increasingly intricate human-AI interplay.

Group Decision and Negotiation
Gain

A systematic review and bibliometric analysis of 627 articles produced a conceptual framework with two dimensions and four paradigms that advances theoretical understanding of hybrid decision-making and provides actionable insights for organizations in AI-driven environments.

Group Decision and Negotiation
Beyond bias: using AI to reduce diagnostic noise and manage novelty in clinical reasoning
HealthContested · G 71 / P 70

AI impact on diagnostic accuracy in clinical reasoning

Source article: Beyond bias: using AI to reduce diagnostic noise and manage novelty in clinical reasoning

Problem

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

Diagnosis
Gain

AI reduces unwanted variability in structured pattern recognition tasks such as imaging analysis, improving diagnostic consistency and mitigating noise-driven errors.

Diagnosis
The rapid rise of housefishing: are AI-enhanced property listings helpful – or sinister?
LifestyleContested · G 53 / P 54

AI-enhanced and AI-staged property listing photos shown to prospective homebuyers

Source article: The rapid rise of housefishing: are AI-enhanced property listings helpful – or sinister?

Problem

AI-enhanced listings misrepresent size, condition and layout, leading buyers to waste time viewing homes that do not match the photos.

The Guardian
Gain

AI programs let estate agents quickly and cheaply stage and enhance listing photos to help buyers visualise a property's potential.

The Guardian
Is the most popular song played on Australian radio stations the product of generative AI?
Media & ArtsContested · G 55 / P 52

AI-generated cover of Like a Prayer achieving No 1 radio airplay and commercial streaming success

Source article: Is the most popular song played on Australian radio stations the product of generative AI?

Problem

AI-generated cover topping radio airplay raises worrying questions about valuing human expression and diverts streaming and radio royalties while training on artists' work.

The Guardian
Gain

Producer used generative AI as a tool to create a cover that provided listeners with good music and achieved 35m Spotify streams and No 1 on the National Radio Airplay chart.

The Guardian