# AesBiasBench: Evaluating Bias and Alignment in Multimodal Language Models for Personalized Image Aesthetic Assessment

arXiv: https://arxiv.org/abs/2509.11620 · HTML: https://mlsanity.com/paper/2509.11620/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- Placed in Image editing and quality assessment, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 25 of its 40 references that are in the map.
- First posted on arXiv 15 Sep 2025.

## Placed in

- Multimodal LLMs / vision-language models / [Image editing and quality assessment](https://mlsanity.com/mllm/image-editing-and-quality-assessment/)
- LLM alignment / RLHF / preference optimisation / [Diffusion and multimodal model alignment](https://mlsanity.com/alignment/diffusion-and-multimodal-model-alignment/)

## Builds on (most cited first, 25 of 25)

- [Improved Baselines with Visual Instruction Tuning](https://mlsanity.com/paper/2310.03744/) (2023)
- [ChatGLM: A Family of Large Language Models from GLM-130B to GLM-4 All Tools](https://mlsanity.com/paper/2406.12793/) (2024)
- [How Far Are We to GPT-4V? Closing the Gap to Commercial Multimodal Models with Open-Source Suites](https://mlsanity.com/paper/2404.16821/) (2024)
- [DeepSeek-VL2: Mixture-of-Experts Vision-Language Models for Advanced Multimodal Understanding](https://mlsanity.com/paper/2412.10302/) (2024)
- [BOLD: Dataset and Metrics for Measuring Biases in Open-Ended Language Generation](https://mlsanity.com/paper/2101.11718/) (2021)
- [Editing Large Language Models: Problems, Methods, and Opportunities](https://mlsanity.com/paper/2305.13172/) (2023)
- [MLLM-as-a-Judge: Assessing Multimodal LLM-as-a-Judge with Vision-Language Benchmark](https://mlsanity.com/paper/2402.04788/) (2024)
- [Do LLMs exhibit human-like response biases? A case study in survey design](https://mlsanity.com/paper/2311.04076/) (2023)
- [Evaluating and Mitigating Discrimination in Language Model Decisions](https://mlsanity.com/paper/2312.03689/) (2023)
- [It's All in the Name: Mitigating Gender Bias with Name-Based Counterfactual Data Substitution](https://mlsanity.com/paper/1909.00871/) (2019)
- [Q-Insight: Understanding Image Quality via Visual Reinforcement Learning](https://mlsanity.com/paper/2503.22679/) (2025)
- [Mono-InternVL: Pushing the Boundaries of Monolithic Multimodal Large Language Models with Endogenous Visual Pre-training](https://mlsanity.com/paper/2410.08202/) (2024)
- [AesBench: An Expert Benchmark for Multimodal Large Language Models on Image Aesthetics Perception](https://mlsanity.com/paper/2401.08276/) (2024)
- [Personalized Image Aesthetics Assessment with Rich Attributes](https://mlsanity.com/paper/2203.16754/) (2022)
- [Investigating Bias in LLM-Based Bias Detection: Disparities between LLMs and Human Perception](https://mlsanity.com/paper/2403.14896/) (2024)
- [AesExpert: Towards Multi-modality Foundation Model for Image Aesthetics Perception](https://mlsanity.com/paper/2404.09624/) (2024)
- [Image Aesthetic Assessment: An Experimental Survey](https://mlsanity.com/paper/1610.00838/) (2016)
- [Towards Artistic Image Aesthetics Assessment: a Large-scale Dataset and a New Method](https://mlsanity.com/paper/2303.15166/) (2023)
- [UNIAA: A Unified Multi-modal Image Aesthetic Assessment Baseline and Benchmark](https://mlsanity.com/paper/2404.09619/) (2024)
- [Balancing the Picture: Debiasing Vision-Language Datasets with Synthetic Contrast Sets](https://mlsanity.com/paper/2305.15407/) (2023)
- [textttModSCAN: Measuring Stereotypical Bias in Large Vision-Language Models from Vision and Language Modalities](https://mlsanity.com/paper/2410.06967/) (2024)
- [LAPIS: A novel dataset for personalized image aesthetic assessment](https://mlsanity.com/paper/2504.07670/) (2025)
- [LLaVA-SpaceSGG: Visual Instruct Tuning for Open-vocabulary Scene Graph Generation with Enhanced Spatial Relations](https://mlsanity.com/paper/2412.06322/) (2024)
- [ICM-Assistant: Instruction-tuning Multimodal Large Language Models for Rule-based Explainable Image Content Moderation](https://mlsanity.com/paper/2412.18216/) (2024)
- [Mitigating Social Biases in Language Models through Unlearning](https://mlsanity.com/paper/2406.13551/) (2024)

## Built on by (most cited first, 11 of 11)

- [Vesta: A Generalist Embodied Reasoning Model](https://mlsanity.com/paper/2606.20905/) (2026)
- [Can MLLMs Critique Like Humans? Evaluating Open-Ended Aesthetic Reasoning in Multimodal Large Language Models](https://mlsanity.com/paper/2606.29689/) (2026)
- [Position: Universal Aesthetic Alignment Narrows Artistic Expression](https://mlsanity.com/paper/2512.11883/) (2025)
- [The Algorithmic Gaze of Image Quality Assessment: An Audit and Trace Ethnography of the LAION-Aesthetics Predictor](https://mlsanity.com/paper/2601.09896/) (2026)
- [StylisticBias: A Few Human Visual Cues Drive Most Social Biases in MLLMs](https://mlsanity.com/paper/2606.20527/) (2026)
- [Efficient Personalization of Generative User Interfaces](https://mlsanity.com/paper/2604.09876/) (2026)
- [Interoceptive Divergence in Aesthetic Evaluation and Implications for Human-AI Alignment](https://mlsanity.com/paper/2605.18759/) (2026)
- [Symphony of Bias: Exploring Gender Associations with Musical Instruments in Multimodal LLMs](https://mlsanity.com/paper/2607.26355/) (2026)
- [The Safeguard Worked. Is the LLM System Safer?](https://mlsanity.com/paper/2609.00519/) (2026)
- [Learning Heterogeneous Preferences](https://mlsanity.com/paper/2609.17847/) (2026)
- [Understanding Machine Taste: Aesthetic Agreement Across Foundation Models in K-pop](https://mlsanity.com/paper/2610.17731/) (2026)

## Abstract (arXiv)

Multimodal Large Language Models (MLLMs) are increasingly applied in Personalized Image Aesthetic Assessment (PIAA) as a scalable alternative to expert evaluations. However, their predictions may reflect subtle biases influenced by demographic factors such as gender, age, and education. In this work, we propose AesBiasBench, a benchmark designed to evaluate MLLMs along two complementary dimensions: (1) stereotype bias, quantified by measuring variations in aesthetic evaluations across demographic groups; and (2) alignment between model outputs and genuine human aesthetic preferences. Our benchmark covers three subtasks (Aesthetic Perception, Assessment, Empathy) and introduces structured metrics (IFD, NRD, AAS) to assess both bias and alignment. We evaluate 19 MLLMs, including proprietary models (e.g., GPT-4o, Claude-3.5-Sonnet) and open-source models (e.g., InternVL-2.5, Qwen2.5-VL). Results indicate that smaller models exhibit stronger stereotype biases, whereas larger models align more closely with human preferences. Incorporating identity information often exacerbates bias, particularly in emotional judgments. These findings underscore the importance of identity-aware evaluation frameworks in subjective vision-language tasks.