Bias in Vision-Language Models: Income Inequality in Contrastive Language-Image Pre-training (CLIP)

Biases in vision-language models are increasing the digital divide. Understanding these biases is key to making artificial intelligence more equitable.

Natural language processing (NLP) and computer vision (CV) are some of the most impactful areas of artificial intelligence. However, while their current models perform well overall, performance reports normally exclude details about how the models perform on particular groups within the data. Specifically, data collected from low-income households tend to be overlooked during model evaluation, which causes the AI to …

artificial artificial intelligence bias biases clip computer computer vision current data science digital ethics however image income inequality intelligence key language language models language processing machine learning making natural natural language natural language processing nlp normally performance pre-training processing reports training understanding vision vision-language models