View

The Study

Human versus artificial intelligence: investigating ability of young academics from research and non-research institutions to identify ChatGPT-generated dental research abstracts

In simple terms

This study watched six new dental researchers try to guess which abstracts were written by a robot and which were written by people. It found that sometimes they got it right and sometimes they didn’t—and the robot detectors were better at it. But it didn’t prove that robots are always better or that people always fail—it just showed what happened in this one small test.

44%

Analysis score

44/ 44

Maximum 44 for a cross-sectional study.

Where the score came from

Reporting40
Methodology68
Publication100
Statistical54
Study type (basis of the score)
Cross-Sectional Study
Level 4 - Case series
What’s the bottom line?

Scientists gave young dentists 150 abstracts—half written by humans, half by ChatGPT—and asked them to guess which was which. They also tested computer tools to see which could spot the robot-written ones.

Where does this study sit?

Reviews of RCTs (Meta-analyses)

Max 100

Randomized Trials

Max 90

Reviews of Cohort Studies

Max 85

Cohort Studies

Max 72

Reviews of Case-Control Studies

Max 63

Case-Control Studies

Max 58

Cross-Sectional & Case Series

Max 50

Expert Opinion

Max 5
StrongerWeaker
Cross-Sectional & Case Series
Level 4
44

44 / 100

Quality score

Snapshots of a population at a single point in time, or descriptions of small groups. Can identify correlations and prevalence, but cannot determine cause and effect.

Cannot establish causation

Save studies & get personalized insights

Create a free account to save this study, track new evidence as it comes in, and get breakdowns of studies in the topics you care about.

Key takeaways

Summary

Based on the study abstract and findings.

  1. 1Yes—this means even trained young researchers can't reliably spot AI-written science, but computers can, especially Turnitin and GPTZero.
  2. 2Humans got it right only 44% to 76% of the time.
  3. 3GPTZero got it right 90% of the time.
  4. 4Turnitin got it right 94% of the time.
  5. 5Robot abstracts were rated lower quality and looked very different from human ones.

Score breakdown, methodology, conflicts of interest, evidence analysis & raw study data

Publication

Journal

Scientific Reports

Year

2026

Authors

Matheel Al-Rawas, Omar Abdul Jabbar Abdul Qader, G. S. S. Lin, Yew Hin Beh, Muhammad Annurdin Sabarudin, Yee Ang, J. Low, J. Abdullah, T. Noorani

Open Access
Analysis v5
Fit Body Science verdict — we translate health studies into clear verdicts backed by peer-reviewed research.

Not medical advice. For informational purposes only. Always consult a qualified healthcare professional before making health decisions.