Google, like AWS and Azure, has very powerful machine learning APIs that are available for a fee. These ML/AI APIs are pretrained, meaning that you don't need to bother writing any ML models of your own; you can just use the great models Google has already built. Use cases include the detection of explicit content, celebrities, and other specific features within video. Notice that scanning 10,000 minutes of video for a few different types of features costs about $2,250/month, as shown in the following screenshot:
