TruaceTracing the truth around AIThursday, August 27, 2026
The Index

What the evidence says.What the public feels.

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

1,169 results
Show filters and sorting

AI gains · 649

68
GainEducation· Stable· Evidence: Moderate (1 source)

ChatGPT use was associated with meaningful improvement in academic writing for university students in the UAE, especially for those using English as a second language.

A peer-reviewed study published December 26, 2025 surveyed 506 students from several higher education institutions in the United Arab Emirates about ChatGPT use. It found usage varied by gender, academic level, and field of study, with female students, senior students, and non-technical program students reporting higher or more writing-focused use, and concluded the tool improves academic writing for ESL students.

Impact 30%49
Evidence 25%95
Scale 20%35
Confidence 15%87
Recency 10%93

Updated Jul 20, 2026 · TRV-2026-0425

68
GainEducation· Stable· Evidence: Moderate (1 source)

In EFL education, higher digital literacy and trust in AI increased perceived ease of use and usefulness, strengthened attitudes and intentions to continue using AI, and supported creative language learning.

By December 2025, researchers surveyed EFL students in two stages to understand AI acceptance. The first survey of 460 students linked digital literacy to higher perceived ease of use and usefulness and greater trust in AI, which in turn shaped attitudes and intentions to use AI for language learning. The second survey of 640 students split trust into Human-like Trust and Functionality Trust and found differentiated effects on continued use and relational engagement.

Impact 30%49
Evidence 25%95
Scale 20%35
Confidence 15%87
Recency 10%93

Updated Jul 20, 2026 · TRV-2026-0424

68
GainEducation· Stable· Evidence: Moderate (1 source)

Systematic review of 40 studies reports AI tools for personalized learning, tutoring and assessment improved student engagement, learning performance and teaching efficiency through adaptive feedback and real-time analytics in bi/multilingual settings.

A December 2025 systematic review of 40 articles from 2015-2025 examined how AI is used in bi/multilingual education, focusing on personalized learning, intelligent tutoring systems and chatbots, and automated assessment. It reported that adaptive feedback and real-time analytics were associated with higher student engagement, learning performance and teaching efficiency in multiliteracy language learning.

Impact 30%49
Evidence 25%95
Scale 20%35
Confidence 15%87
Recency 10%93

Updated Jul 20, 2026 · TRV-2026-0423

68
GainEducation· Stable· Evidence: Moderate (1 source)

Google Gemini can process text, image, audio, and video inputs to generate diverse content types for educational use.

Published May 23, 2024, this emerging technology report reviews Google Gemini as a multimodal generative AI tool, describing its ability to process text, image, audio, and video inputs and generate diverse content, and summarizing recent empirical studies and technology-in-practice examples in education.

Impact 30%49
Evidence 25%95
Scale 20%35
Confidence 15%87
Recency 10%93

Updated Jul 20, 2026 · TRV-2026-0419

AI problems · 520

56
ProblemMedia & Arts· Newly added· Evidence: Moderate (1 source)

The Department of Marine and Coastal Resources drew strong public criticism after its AI art contest for National Dugong Conservation Day selected a winning image that inaccurately depicted a dugong with four legs.

On August 17, 2026, the Bangkok Post reported that the Department of Marine and Coastal Resources faced strong criticism after announcing winners of an AI art contest held for National Dugong Conservation Day. The controversy centered on the winning image, which showed a dugong with what appeared to be four legs.

Impact 30%49
Evidence 25%62
Scale 20%35
Confidence 15%62
Recency 10%98

Updated Aug 18, 2026 · TRV-2026-0816

56
ProblemMedia & Arts· Newly added· Evidence: Moderate (1 source)

Former lead writer Stella Sacco alleges she was replaced by ChatGPT midway through development, after the CEO said he would be happy to replace her with AI.

On August 17, 2026, The Verge reported that Saber CEO Matthew Karch apologized to former Rideshare 'Stimulator' lead writer Stella Sacco after telling This Week in Videogames he would have been 'happy to replace her with AI.' Sacco had claimed Saber replaced her with ChatGPT midway through development, a claim Karch denied, while the game's Steam page states it uses AI for voices, music, and missions.

Impact 30%49
Evidence 25%62
Scale 20%35
Confidence 15%62
Recency 10%98

Updated Aug 18, 2026 · TRV-2026-0814

56
ProblemCrime· Newly added· Evidence: Moderate (1 source)

AI-generated deepfakes impersonating Prime Minister Anthony Albanese and other public figures were used in fake investment pitches, contributing to $7.4m in reported losses to the top 10 impersonated figures.

By 16 August 2026, Australia's corporate watchdog ASIC warned of a steep rise in investment scams using AI-generated deepfakes, most commonly of Prime Minister Anthony Albanese. One circulating video used real footage overlaid with fake audio promising $4,000 could earn $40,000 a month as an 'official platform' with a government guarantee. Reports to Scamwatch linked $7.4m in losses to the top 10 impersonated public figures.

Impact 30%49
Evidence 25%62
Scale 20%35
Confidence 15%62
Recency 10%98

Updated Aug 17, 2026 · TRV-2026-0793

56
ProblemMedia & Arts· Newly added· Evidence: Moderate (1 source)

Removing the visible watermark makes AI-generated images, videos and music less immediately identifiable as synthetic to casual viewers, raising risk of audience confusion about authenticity.

On August 15, 2026, Google changed Gemini to let users remove visible watermarks from AI-generated images, videos and music. The system continues to embed invisible provenance through SynthID and C2PA data even when the visible mark is turned off.

Impact 30%49
Evidence 25%62
Scale 20%35
Confidence 15%62
Recency 10%98

Updated Aug 16, 2026 · TRV-2026-0776

Recomputed live from the record · Aug 27, 2026, 8:34 PM