How are museums using artificial intelligence, and is AI the future of museums?

This article does a good job of exploring the multiple different avenues through which museums are being affected by AI. While many headlines typically focus on the use of AI in exhibits this article also delves into the important topic of how AI is being used to improve museums as a whole on the service and logistics side. One of the examples the article went into along this front was the humanoid robot Pepper used in the Smithsonian, which helps visitors with their questions and even poses for photos. In the future, I could see AI being used to make information much more accessible to visitors, both about museums and exhibits. Even with our current level of technology, it would be possible for museums to train large language models to answer questions about different exhibits and amenities so visitors could essentially always have a personal guide at their fingertips.

A MUSEUM’S EXPERIENCE WITH AI

In this article, the CHM takes the reader through a deep dive into how AI was specifically incorporated into an effort at the CHM to make collections more easily accessible and searchable. The article delves into the incredible amount of development work the CHM team undertook by collaborating with external partners and utilizing Microsoft’s Cognitive Services. In all the team was able to make a successful prototype of a portal containing digital assets and AI-generated metadata. Although the team originally had much grander ideas in mind about creating a sort of graph view to show the connection between search results, they eventually decided to put that aside to limit the scope of their prototype. The results of the evaluation showed that external evaluators were more positive about the utility of the machine-generated metadata than internal evaluators, and they found that automatic transcription of video and audio items using machine-learning tools seemed to be the most useful feature. In all this article does a good job of illustrating all the difficult development and design work that goes into building AI-enabled services, and the great potential that these services can have.