Towards Understanding the Use of MLLM-Enabled Applications for Visual Interpretation by Blind and Low Vision People