Article Per Year (5 Year)
p-Index From 2021 - 2026
0.23
P-Index
This Author published in this journals
All Journal MIX : Jurnal Ilmiah Manajemen
Andreas Rumata Simanjuntak
Industrial Chemical Technology Polytechnic
Decision Sciences, Operations Research & Management Economics, Econometrics & Finance Social Sciences

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Comparative Study of Artificial Intelligence Accuracy on Project Team Member Assessment Poltak Evencus Hutajulu; Andreas Rumata Simanjuntak; Toba Sastrawan Manik; R Reza El Akbar
MIX: JURNAL ILMIAH MANAJEMEN Vol. 16 No. 2 (2026): MIX : Jurnal Ilmiah Manajemen
Publisher : Universitas Mercu Buana

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.22441/jurnal_mix.2026.v16i2.014

Abstract

Objectives: This study compares five AI platforms (ChatGPT 5.2, Claude Sonnet 4.5, Gemini 3 Pro, GLM 4.7, DeepSeek V3.2) with manual HR assessment in team performance evaluation, examining accuracy and predictive validity by correlating with program success rates.Methodology: Using a within-subjects design, 32 team members managing 619 locations were evaluated by six methods via standardized prompts. Analysis included comparative tests, accuracy metrics, and correlation between performance scores and graduation rates.Finding: All AI platforms scored higher than the HR baseline. Correlation analysis revealed substantial differences in predictive validity. Claude Sonnet 4.5 demonstrated the strongest correlation with graduation rates (ρ = 0.70), followed by GLM 4.7 (ρ = 0.67), accounting for approximately 50-55% of the outcome variability. The HR baseline unexpectedly showed a negative correlation with program success (ρ = -0.49), indicating misalignment between the evaluation criteria and actual effectiveness. DeepSeek V3.2 most closely matched HR scores but showed only moderate predictive validity, while ChatGPT 5.2 showed no meaningful relationship with outcomes.Conclusion: AI platforms differ substantially in predictive validity. Organizations should prioritize outcome-based validation over human agreement when selecting AI for HR decisions and consider recalibrating manual evaluation systems.
Co-Authors Hutajulu, Poltak Evencus R Reza El Akbar Toba Sastrawan Manik